Validation of the Dynamic Simulation Capabilities of Serpent2/Subchanflow Using Experimental Data from the Research Reactor SPERT IV D-12/25

Research nuclear reactors are critical to the development of nuclear technology, but because of the complex configuration of the fuel assemblies and the different initial operating conditions, neutronic and thermal-hydraulic analyzes are performed with dedicated or adapted codes. These simulation tools might not be as accurate, since they use simple resolution methods and assumptions that do not always capture all aspects of the behavior of a nuclear research reactor. On the other hand, as time goes by, the evolution of numerical tools for the core analysis of power reactors have experienced a considerable progress. Nowadays, pin/subchannel level analysis of the core with coupled codes based on transport (SP3, MOC, SN, etc.) or Monte Carlo methods are applied in addition to the nodal diffusion codes. Hence, the research community is adapting and validating selected high-fidelity tools developed for power reactors to perform detail core analysis of e.g. Material Testing Reactors (MTR) cores at plate and subchannel level. This work deals with the validation of the high-fidelity coupled Serpent2/Subchanflow, which was modified and extended for the plate/subchannel analysis of MTR-cores, using the data of rod ejection tests performed in the SPERT IV D-12/25 reactor, especially the tests B-34 and B-35 were selected to validate the dynamic capability of Serpent2/Subchanflow. In these unique tests, experimental data e.g. thermal neutron flux, core power evolution during the rod ejection tests, and the plate cladding temperature was measured. It is noted that, due to the lack of detailed information on the initial conditions, the extraction and introduction scenarios of the transient rod for the reactivity insertion required calibrations and assumptions regarding velocity and position. The comparison of selected parameters predicted by the coupled simulations at plate/subchannel level of the SPERT IV reactor with the measured data at static and transient conditions shows excellent agreement confirming the high accuracy appropriateness of the used code for the analysis of research reactors. The calculated values of thermal neutron flux and core power evolution have a statistical error of +/- 2 sigma. It was also found that the maximum temperature difference between calculated and experimental values is 7 C-degrees and similar to 10 C-degrees for tests B-34 and B-35, respectively. In addition, the coupled code predicts for the first time the temperature of each plate and subchannel considering the local feedbacks between neutronics and thermal-hydraulics allowing the identification of the hottest/coldest plate in the core. The high-fidelity validation tool can provide comparison solutions for current research reactor core analysis methods, such as core analysis with point kinetics or 3D nodal diffusion codes coupled with fuel assembly-level thermal-hydraulics.