An Efficient And Reliable Distributed Data Acquisition System Based On Industrial Cloud

The arrival of intelligent era,which means that the arrival of industrial big data,industrial applications of big data will become the basis for future competition and growth of industrial enterprises,and is the key to achieving intelligent industrial growth.The quality of industrial data collection is the precondition for the construction of industrial big data platform and the following series of applications in industrial big data environment.Therefore,it is of great significance to study how to collect large-scale distributed data efficiently and reliably.Because the sampling frequency of MES system and monitoring station is low,the sampling requirements of the host computer is not very strict,and the actual data acquisition cycle is larger than the set acquisition cycle.In the big data industry environment,we often need to the analysis of the data acquisition cycle real-time correlation,so sequential data have more strict,efficiency and reliability of the existing data acquisition system is not enough to meet the requirements.In order to improve the reliability of data collection,important data corresponding to the data acquisition server is usually 1:1 redundancy configuration,which will lead to data acquisition server resource utilization low.General data corresponding to the data acquisition server failure will lead to the loss of the corresponding data,data acquisition server reliability is low.There are often some data sets or data items are missing in current data acquisition systems.In the process of industrial monitoring,it is necessary to monitor thousands or even tens of thousands of data points.Moreover,the real-time requirements of data acquisition and processing are high,which usually require the acquisition and processing of monitoring data in seconds or even milliseconds.The arrival of large industrial data,industrial process data collection of real-time put forward higher requirements.With the production of enterprises running,increasing the amount of data storage enterprises and historical data for the importance of enterprise analysis,data compression put forward higher requirements.To solve these problems,this paper relies on Northeastern University,State Key Laboratory of Integrated Automation of Process Industry to carry out research,highly efficient and reliable distributed data acquisition system based on industrial cloud.The main research work is as follows:Firstly,research on the existing data acquisition system at home and abroad and survey on the production of the enterprise distributed data acquisition.Based on this,combined with the research needs of the laboratory and production needs of enterprises,highly efficient and reliable distributed data acquisition system based on industrial cloud is proposed.At the same time the functional requirements of the system and performance requirements were analyzed in detail.Secondly,the system is designed according to the requirement analysis.According to the modular design principle,the system architecture design,software architecture design and software function module design are completed.The system consists of industrial cloud platform layer,control layer,equipment layer and industrial site composition.The industrial cloud platform layer comprises a management node module,a task distribution node module,a task storage module,a collection node module and a database module.Task node module,task warehouse module,collection node module and database module are designed respectively.Thirdly,according to the design scheme,based on the industrial cloud platform,the development of the system is completed by Java development language,SQL Server 2012 database and Fourinone framework.The system performs the task migration on the basis of the consistent hash algorithm,so that the resource utilization rate of the acquisition nodes reaches relatively balanced.The data compression method proposed in this paper can compress the numerical data and the Boolean data separately and achieve a higher compression ratio,saving data storage space greatly.The system can create,modify and delete collection tasks according to users' needs at any time,and can set different priorities and sampling periods for different acquisition tasks according to actual needs.The ta sk assignment node automatically assigns the tasks according to the detected operation information of the acquisition nodes and the tasks set by the user,so that the acquisition nodes achieves a relatively balanced load.The system allows any acquisition node to join and offline to improve the scalability of the acquisition node.When some collection node failure,the collection of tasks between the nodes can be migrated to other acquisition node to improve the reliability of data collection.The acquisition nodes can adopt diflferent configurations to improve the system's ease of use.Lastly,the system is deployed in the Industrial Cloud Computing Center,and the data simulation platform of the wind turbines is built with MatrikonOPC.The ninth function,which contains task management,task allocation,data acquisition,data compression and the reliability,high efficiency and expansibility of the system reach validation results.Verifying the reliability from the overall system,the nodes and data groups and data items.The high efficiency verification of this system performs from collection and high storage.Verifying the scalability from nodes,data groups and data items.Experimental results show that the system has good reliability,high efficiency and scalability in the process of large-scale distributed data acquisition,and greatly reduces the data storage space.This system can meet the requirements of industrial enterprises to collect data efficiently and reliably.The data compression method proposed in this paper greatly reduces the data storage space.The system also requires more support and integration of images and video.