Design and implementation of a relational database of macroscopic cross-sections for steady-state and transient analysis in a medium sized PWR

摘要

This paper deals with the design and implementation of a relational database of macroscopic cross-sections as well as validation and its usage as a component in steady-state and transient analysis. These types of analysis are definitely designated as data-driven processes in which their behavior is changed based on data. The data in this manner can be the variation of fuel composition density as well as the fuel temperature during the reactor operation or in some accidents. In order to gain the accurate results in transient analysis, applying these variations is inevitable. Therefore, for each changing in fuel behavior, the new set of macroscopic cross-sections must be reproduced by performing the lattice calculations at run-time. During recent years some approaches have been done in both academic and commercial areas to take advantage of already calculated cross-sections instead of recalculate them by lattice codes during the analysis. Using this approach leverages the computation performance by reducing the floating point operations which can decrease the CPU load as well as the calculation time. Although this approach delivers a good performance in terms of time and computation, when a large number of fuel or reactor core circumstances are subjected to be analyzed the accuracy is became an important point. In this situation the cross-sections for all expected conditions must be already calculated. This cause the large number of cross-section data which must be stored and managed somehow. The system applied in this study to store and manage these cross-section data sets is a relational database management system (RDBMS) which is based on Microsoft SQL Server 2008 Management System. The node wise macroscopic cross-sections for each nominal and off-nominal of a typical medium sized PWR reactor core conditions which are calculated by WIMS-D5 and CITATION codes are tabulated in this system and can be retrieved by structured query language extremely rapid on demand.