Development and Preliminary Verification of a Neutronics-Thermal Hydraulics Coupling Code for Research Reactors with Unstructured Meshes

Mingrui Yang, Chixu Luo, Dan Wang, Tianxiong Wang, Xiaojing Liu, Tengfei ZhangSpringer Proceedings in Physics 2023 · DOI · OpenAlex

Abstract

To maximize their adaptability and versatility, research reactors are designed to adapt to various operational conditions, resulting in complex configurations and irregular geometries. A three-dimensional neutronics-thermal hydraulics (N-TH) coupling code, CENTUM (CodE for N-Th coupling with Unstructured Mesh), has been developed to address these challenges. Steady-state and transient neutronic analyses are performed using a 3D triangular-z nodal transport solver integrated with the stiffness confinement method (SCM), while thermal-hydraulics calculations utilize a multi-channel model. Preliminary verification using TWIGL, 3D-LMW, and NEACRP benchmarks demonstrates high accuracy, with maximum power errors within -1.27% for TWIGL and 3.84% for 3D-LMW. For the NEACRP benchmark, CENTUM showed excellent agreement with reference solutions, achieving a maximum relative power deviation of only 0.055% during transient analysis.

Introduction

Research reactors serve diverse roles, including medical isotope production and material testing, necessitating highly irregular geometries and complex control rod configurations. Traditional simulation codes often struggle with these non-standard shapes due to their reliance on rigid Cartesian or hexagonal grids. Furthermore, the physical phenomena within these reactors are strongly coupled: fission-generated heat alters material temperatures, which subsequently modifies neutron cross-sections. Accurately simulating these interactions during rapid power transients is computationally demanding due to the "stiffness" of the underlying equations.

Methodology

The CENTUM code was developed to provide a robust framework for simulating these complex environments through several key technical innovations.

Unstructured Meshing and Transport Solver

CENTUM employs a 3D triangular-z nodal transport solver. Unlike traditional rectangular grids, this unstructured approach allows the mesh to conform precisely to the irregular boundaries and internal structures characteristic of research reactors. This geometric flexibility is essential for modeling the heterogeneous cores of modern research facilities.

Triangular mesh geometry used in the CENTUM nodal transport solver.

Stiffness Confinement Method (SCM)

To handle the temporal "stiffness" of neutron kinetics equations, CENTUM utilizes the Stiffness Confinement Method. SCM effectively decouples the rapidly changing neutron flux from the slower-evolving delayed neutron precursor concentrations. This allows the code to maintain numerical stability and accuracy even when using relatively large time steps, significantly enhancing computational efficiency during transient simulations.

Neutronics-Thermal Hydraulics Coupling

The integration of neutronics and thermal-hydraulics is managed via the Operator Splitting Semi-Implicit (OSSI) method. In this scheme, the power distribution calculated by the neutronics module is passed to the thermal-hydraulics module to update temperature fields. These updated temperatures are then used to recalculate cross-sections for the subsequent neutronics step, ensuring a consistent feedback loop throughout the simulation.

Schematic of the OSSI coupling between neutronics and thermal-hydraulics modules over successive time steps.

Verification and Results

The code was verified against three industry-standard benchmarks to ensure both spatial and temporal accuracy:

  • TWIGL: A 2D kinetics problem where CENTUM achieved maximum power errors of -1.27% and -0.45% for different perturbation cases.
  • 3D-LMW: A spatial-temporal benchmark involving control rod movement. CENTUM showed a 3.84% relative power error compared to the SPANDEX reference, an acceptable result given the transition from diffusion to transport modeling.
  • NEACRP: A coupled N-TH benchmark. CENTUM predicted critical boron concentration within 0.35 ppm and fuel Doppler temperature within 2.16 °C of reference values.

Discussion and Conclusion

A significant finding of this work is that the SCM approach provides a stable "amplitude frequency" for power calculations; even if minor oscillations occur at larger time steps, the average values remain highly accurate. This allows for faster simulations without compromising the physical integrity of the results. While current benchmarks utilize Cartesian geometries, the underlying triangular-z solver positions CENTUM as a powerful tool for future analysis of non-Cartesian research reactor designs. Future work will focus on validating the code against experimental data from operating research reactors and further optimizing the thermal-hydraulics multi-channel models.

Summary updated

Abstract

Abstract To maximize their adaptability and versatility, research reactors are designed to adapt to various operational conditions. These requirements result in more complex configurations and irregular geometries for research reactors. Besides, there is usually a strong coupling of neutronics-thermal hydraulics (N-TH) fields inside the reactor. A three-dimensional N-TH coupling code has been developed named CENTUM (CodE for N-Th coupling with Unstructured Mesh). Steady-state and transient neutronic analyses are performed using a 3D triangular-z nodal transport solver with the stiffness confinement method (SCM). Meanwhile, thermal-hydraulics calculations adopt a multi-channel model. For a preliminary verification of the code, we examine CENTUM with benchmark problems including TWIGL, 3D-LMW, and NEACRP. CENTUM produces maximum power errors of −1.27% and −0.45% for the TWIGL A1 and A2 cases, respectively. For the 3D-LMW benchmark, the largest relative power error of 3.84% is observed at 10 s compared with the reference SPANDEX code. For the NEACRP N-TH coupling benchmark, CENTUM results in a 0.35 ppm error in critical boron concentration, a 2.16 °C discrepancy in the fuel average Doppler temperature, and a 0.63% overestimation in the maximum axial power. Moreover, transient results considering thermal-hydraulics feedback are in good agreement with the PARCS reference solutions, with the maximum relative power deviation being only 0.055%.
BibTeX
@inproceedings{yang2023development,
  author = {Mingrui Yang and Chixu Luo and Dan Wang and Tianxiong Wang and Xiaojing Liu and Tengfei Zhang},
  title = {{Development and Preliminary Verification of a Neutronics-Thermal Hydraulics Coupling Code for Research Reactors with Unstructured Meshes}},
  booktitle = {Springer Proceedings in Physics},
  year = {2023},
  doi = {10.1007/978-981-99-1023-6\_58},
}
Figure