RMC-based computational study of CABRI-like benchmark

Background Reactivity-initiated accidents (RIAs) are critical safety concerns for pressurized water reactors. The CABRI research reactor, equipped with a unique He-3 transient rod system, can generate extreme power pulses closely resembling RIA conditions, serving as a key experimental platform for fuel transient behavior studies. The CABRI-like benchmark problem, established under an international framework, is designed to test transient simulation capabilities of numerical codes, with several internationally recognized codes such as TRIPOLI-4® and Serpent participating in cross-comparisons. The domestically developed Monte Carlo code RMC possesses built-in transient kinetics and temperature-feedback modules, offering an independent computational tool for such benchmark analyses.

Purpose This study aims to perform independent neutronics/thermal-hydraulics coupled transient calculations for the CABRI-like benchmark using RMC, and to cross-validate the results against those obtained from international peer codes, thereby assessing the rationality of the benchmark and the consistency of the numerical results.

Methods A detailed CSG (Constructive Solid Geometry) model of the benchmark core was constructed with RMC, incorporating all 40 assemblies including fuel assemblies, BCS control assemblies, and He³ transient assemblies. The LOW transient scenario, driven by a prescribed rapid decrease in He³ density, was simulated for the first 0.15 s, discretized into 150 time steps of 0.001 s each. The RMC space-time kinetics module based on the predictor-corrector quasi-static method was employed. Neutronics/thermal-hydraulics coupling was realized via a Python script that updates the fuel temperature of each assembly at every time step according to the deposited energy, while the built-in Target Motion Sampling (TMS) algorithm provides on-the-fly temperature-dependent cross sections, capturing the Doppler feedback effect. Criticality calculations were performed at the beginning and end states of the LOW, INTER, and HIGH transients for comparison of k_eff and reactivity insertion against TRIPOLI-4® and Serpent results. The fuel temperature reactivity coefficient was evaluated by varying the fuel temperature from 300 K to 1 800 K.

Results Criticality comparisons show a maximum k_eff deviation of less than 0.001 and a reactivity insertion deviation of less than 0.000 4. For the LOW transient, the relative power increases to approximately 6 times the initial 1 MW, and the average fuel temperature rises by about 0.26 K, both in good agreement with the reference data. The computed temperature feedback coefficient is approximately -0.000 02/K, consistent with the compared results. Assembly-level distributions of power, energy deposition, and temperature at t = 0.15 s are presented; the highest temperature (300.32 K) occurs in assembly No. 37, and the three columns of assemblies adjacent to the central experimental zone exhibit above-average temperature increases. A clear correlation between the normalized energy deposition rate and the assembly-averaged temperature is observed.

Conclusions The RMC code successfully reproduces the fast transient and neutronics/thermal-hydraulics coupling behavior of the CABRI-like benchmark, with results in good agreement with those from international Monte Carlo codes, confirming both the capability of RMC for such analyses and the validity of the benchmark specification. Future work will introduce axial discretization of the fuel rods to obtain axial power and temperature profiles, thereby providing more complete reference data for the benchmark.