Research On System Reliability For Reconfigurable Manufacturing Grid

Under the tendency of economic globalization and information globalization,manufacturing grid (MG) has become the research focus in networked manufacturingfield. The final goal of MG is to realize manufacturing resource sharing andcollaborative working, provide the better quality of service (QoS), so as to accomplishtasks submitted by consumers (service requestors).With the deep research in MG, the system based on MG application is becomingmore and more complicated. It is hard for the user to choose the services that meettheir requirements. So it is necessary to provide the customized platform whichmatches the user’s requirements. The reconfiguration concept in MG is to add theavailable service needed into the customized platform, which construct a newfunctional MG system. The user can access the MGS service effectively and quicklythrough the Portal, so that improve the system’s availability.Reliability Management (RM) is fundamental and key research field in MG. Inthe process of system reconfiguration, the function of RM is judging new system’sreliability according to the component pattern, if the reliability can’t meet the user’srequirements; the system is limited to application. The successful of RM is not onlyrelative to service efficiency and availability, but also decisive acceptance to RMGSusers.RM of Reconfigurable Manufacturing Grid System (RMGS) was analyzed in thisdissertation, both the theories research and the practical application are provided, thekey technologies in reconfigurability and RM fields was deeply studied, the contentsinclude reconfigurability, reliability, system performance evaluation and self-adaptivefault-tolerance. The application of the key technologies solved the bottleneck ofRMGS in different level enterprises.In order to realize the key technologies of RMGS, Firstly, the requirements forinformation system in enterprise is analyzed, combination with networkedmanufacturing, the architecture of RMGS and software module management areproposed. The mechanism of reconfigurability, MG software module and pattern andMG Portal are studied in this paper, the result is basis of follow-up systemperformance and reliability analysis.Because of the complexity and the dynamitic of MGS, the RMGS customized bythe user also has unstabitily. If RMGS performance and availability is forecasted bysome methods, then the configuration can be optimized in ahead, and efficiency ofRMGS is improved, the better QoS to user is provided. During the using process, theuser can understand the design resource, information resource, system resources andthe trend of RMGS on resource use, contribute to the establishment of job submissionrules, tracking system optimization and planning for future expansion or resourceredistribution; after using, the RMGS monitoring and evaluation information can beviewed in RMGS performance, the user can optimize the allocation of resources andrealize the optimization distribution, reasonable scheduling of application, to obtain high performance; at the same time, the data obtained in the process provide the basisfor the future upgrade or extension of RMGS system. Therefore, the performanceevaluation and forecast is important to RMGS.With the enhancement of computer function, software package is becomingbigger, architecture is becoming more complicated, the quality of software isbecoming more serious. The quality is the soul of software product. Among thequality assessment indices, such as function, reliability, easy-useful, efficiency,maintenance, transplant etc, the reliability is one of the most important issues. TheRMGS belongs to complicated system, if the RMGS’ quality can’t meet the user’srequirements and produce failure in application, then RMGS should be customizedagain by user and reliability of RMGS is important for application.In general, there are two ways to improve reliability, which is fault-avoid andfault-tolerance. Fault-avoid is to avoid fault, it adopts correct design and controlmethod the remove the fault, but can’t avoid all fault in the process. Fault-tolerance isimportant in exception, failure or breakdown of RMGS. In order to improve RMGSquality and reliability, it is necessary to study the mechanism of self-adaptivefault-tolerance, when fault occurs in customization and optimization, system can beintegrated by appropriate component without manual intervene.Lastly, the main of the research achievement have been tested in ManufacturingGrid experiment platform.The system performance evaluation, reliability andself-adaptive in prototype RMGS was realized. The rationality and feasibility ofRMGS was verified, and got much valuable experience and data, it is fundamental toimporve and perfect the RMGS.