Study Of Big Data Based Manufacturing Operational Monitoring And Analyze Platform

With the fast development of industrialization and information technology, widespread application of digital factory, internet of things and other industrial technologies, continuously improvement of industrial automation, the industrial data storage has experienced an exponential growth. Mining valuable information from the growing volume of data to guide industrial control, improve process and reduce cost has become an important strategy for the future enterprises competition. To construct a base platform which can store and manage massive manufacturing related operational and management data and can support intelligent analyze of data, to maximum dig out valuable information, promote information-driven strategy decisions making and timely improvement of industrial real-time control, improve the comprehensive competitiveness of enterprises, has a far-reaching significance.The main characteristics of industrial data are big volume, multi-sourced, heterogeneous (structured, semi-structured and unstructured data existing at the same time) and growing fast. Traditional data process and analyze technologies which are based on relational database have been difficult to meet the demands of industrial applications. Thus it is a must thing to do to study on new and effective industrial big data analyze platform. This paper is target to design a base platform to storage and manage massive data accumulated during industrial production and enterprise management. Main works done in the paper are as follows:(1) Analyze the procedure of large scale industrial data and study difficulties of industrial big data processing, such as storage and management of multi-sourced, heterogeneous industrial big data, the high cost of data storage and management, and industrial real time demands. Based on functional modular design method, design platform functional modules. Based on UML modeling technology analyze the dependencies between functional modules and software structure.(2) Study on performance requirements of big data processing platform architecture, design the platform architecture. By using master-slave architecture to obtain software scalability, using real-time messaging middleware achieve system communication reliability, and based on Hadoop environment to realize industrial big data storage and analysis.(3) A new software architecture analyze method based on system dynamics simulation theory is proposed. That is to see software as a complex system, analyze factors which influence the system performance and the causal relationship between them, build the system model and describe the relationship between factors by using DYNAMO equations.(4) Introduce the data environment of numerical control machine tool factory. Based on the factory environment set modeling parameters value and simulate the platform runtime performance. Analyze system reliability, bottleneck, and possible software structure optimization method.(5) At the end of the paper conclude the whole work done and discuss future research work to do.