International benchmark study of advanced thermal hydraulic safety analysis codes against measurements on IEA-R1 research reactor

In the framework of the IAEA Coordination Research Project on "Innovative methods in research reactor analysis: Benchmark against experimental data on neutronics and thermal hydraulic computational methods and tools for operation and safety analysis of research reactors" the Brazilian research reactor LEA-RI has been selected as reference facility to perform benchmark calculations for a set of thermal hydraulic codes being widely used by international teams in the field of research reactor (RR) deterministic safety analysis. The goal of the conducted benchmark is to demonstrate the application of innovative reactor analysis tools in the research reactor community, validation of the applied codes and application of the validated codes to perform comprehensive safety analysis of RR. The IEA-R1 is equipped with an Instrumented Fuel Assembly (IFA) which provided measurements for normal operation and loss of flow transient. The measurements comprised coolant and cladding temperatures, reactor power and flow rate. Temperatures are measured at three different radial and axial positions of IFA summing up to 12 measuring points in addition to the coolant inlet and outlet temperatures. The considered benchmark deals with the loss of reactor flow and the subsequent flow reversal from downward forced to upward natural circulation and presents therefore relevant phenomena for the RR safety analysis. The benchmark calculations were performed independently by the participating teams using different thermal hydraulic and safety analysis codes that comprise of CATHARE, RELAP5, MERSAT and PARET. The code RELAP5 was used independently by four of the participating teams and therefore the user effect and its impact on the code results can be characterized. The benchmark results demonstrate that most of the codes have the capability to correctly predict the SS case. However, for the LOFA case the simulation results show discrepancies to the measurement although the majority of the applied codes predict a qualitative correct time evolution of the corresponding transients for the coolant and clad temperatures. It is noted that the peak temperatures and the gradients around them are predicted conservatively. The quantitative assessments of benchmark results indicate different amounts of discrepancy between predictions and measurements ranging between 7% and 20% for peak clad temperatures during LOFA. The comparative prediction capability of the employed codes is addressed by additional code-to-code comparisons based on selected TH parameters that comprise flow rate, pressure drop and heat transfer coefficient during natural circulation phase. (C) 2014 Elsevier B.V. All rights reserved.