Safety Analysis of Research Reactors using Advanced Thermal Hydraulic Codes

This report summarizes the main results of the AECS's team in the IAEA's CRP on the Safety significance of postulated initiating events for various types of research reactors and assessment of analytical tools. The CRP was carried out during the period 2003-2006 with the participation of 10 different countries. The contributions of the different teams dealt with the assessment of computational codes related to the safety analysis of research reactors and application of the selected codes on the Brazilian research reactor IEA-R1 that has been considered as reference reactor during the CRP. During the project implementation the codes PARET, COBRA-RERTR and MERSAT have been tested, modified and verified regarding specific phenomena related to safety analysis of research reactors (RR). Due to deficits and limitations in the physical models and the inadequacy for transient calculation the code COBRA was excluded from the application in the safety analysis of IEA-R1 reactor. The application has been then focused on the codes PARET and MERSAT. Both codes were validated against selected separate and integrated effect tests to qualify them for the application in the safety analysis of research reactors. PARET model of IEA-R1 reactor consists of simple approach to model the fuel element groups in the core, whereas MERSAT model comprises complete and detailed modeling of the primary loop including reactor core, bypass, reactor pool, main pump and coupling valve. The performed analyses covered the modelling of thermal-hydraulic and reactor dynamic (core and system calculations) with no source term evaluation. The simulation focused on the most relevant groups of Postulated Initiating Events including LOFA and RIA. The code analyses include the simulation of steady state (SS) operation at nominal power of 5 MW in addition to two groups of design basis accidents consisting of loss of flow accident (LOFA) and reactivity insertion accident (RIA). The simulation results indicate the ability of both codes to model the expected thermal hydraulic and neutron dynamic phenomena. For the LOFA simulation both codes have predicted the onset of flow reversal. Discrepancy was observed in predicting the maximum clad and fuel temperatures, which refers mainly to the applied correlation to simulate natural convection. The achieved results are comparable with other code results like RELAP and MARS being used by other contributing teams. The results of LOFA and RIA demonstrate the safety features of IEA-R1 under the proposed accidents. In Both cases the thermal hydraulic design limits have been not exceeded (for both design limits of thermal hydraulic instability and critical heat flux). However, for the RIA case the maximum clad temperature exceeds the limit being chosen as operation criteria for limiting corrosion effects on the surface of fuel elements. (author)