The Research And Application Of Concurrent Multi-Tasks Technology In Real-Time Monitoring System Based On Multi-Core

SCADA (Supervisory Control And Data Acquisition) systems combinate the advanced computer technology, industrial control technology, display technology and communication technology, measurement technology with itself to complete the function of data acquisition and monitoring. And items detection is one of the typical applications of SCADA. It completes the items’ data acquisition and processing concurrently in real-time, with the characteristics of real-time and concurrent multi-task. So how to high performance, high stability to realize the required functions and high efficiency development become the key factors of developping this kind of system.This paper aims to use the latest technology of multi-core (multi-CPU) computer, transplants an items detection system to multi-core platform, make full use of the advantages of parallism and concurrent of multi-core platform to set up the general architecture of items detection system base on multi-core platform.First of all, do function demand modeling according to a practical items detection software system, then use UML modeling static model of system, expecially do the dynamic modol of system according to the characteristics of multi-task of data acquisition, detection, controlling and executive at the same time, describe the demand of real-time and concurrent multi-task in the system. Base on this, use the technology of object-oriented to design system, the device classes which designed according to the system’s hardware equipment variety and the change of modes segregate the coupling of the actual hardware and system. Next is the core technology of the research in this paper, using parallel tasks supported by task parallel library (TPL) under the.NET4.0to replace original multi-thread technology and to realize concurrent tasks algorithm base on multi-core platform in the detection system; together with using producer-consumer model to realize the communication between different parallel tasks, exchange data and synchronizate tasks. Finally, discusses the possibilities of complex multi-thread multi-task program test and visualization using concurrent visualization of parser in Visual Studio2010, and do a large amount of numerical experiments and analysis and compare, verify the system performance supported by parallel tasks model architecture base on multi-core comparing with the original multi-thread mode improved significantly, provide experience for future improvement of this kind of system and technical support.