Research advances in high-fidelity neutronics computational methods and applications for advanced nuclear reactors

Nuclear reactor technology is of strategic importance to national energy security and defense capabilities. The design and safety analysis of reactor cores rely heavily on neutronics analysis. With the continuous advancement of reactor technology, traditional neutronics calculation methods based on various approximations are increasingly inadequate to meet the precision requirements of advanced reactor analysis. For over two decades, the Nuclear Engineering Computational Physics Laboratory (NECP Lab.) at Xi’an Jiaotong University has conducted systematic research and development on high-fidelity computational methods characterized by “high-precision, high-resolution, and high-confidence”, aiming to address the needs of advanced nuclear reactor development.This paper comprehensively reviews the NECP Lab’s progresses in key theoretical methods for high-fidelity neutronics analysis. NECP lab has developed advanced methods for producing reactor-specific nuclear cross-section libraries. This includes novel approaches for evaluating thermal neutron scattering law data for crystalline and liquid materials using first-principles calculations, overcoming limitations of traditional evaluations. Furthermore, methods for nuclear data uncertainty analysis, based on sampling techniques, and data assimilation, utilizing generalized least squares and Bayesian inference, have been established to quantify and reduce computational biases.Significant innovations have been made in resonance self-shielding calculations, which are critical for accurately treating the sharp variations in neutron cross-sections of nuclides like uranium and plutonium. The team has developed a comprehensive resonance calculation system applicable to various reactor spectra. Specific methods address challenges in different reactor types: for commercial pressurized water reactors (PWRs), global-local coupling methods and equivalent geometry techniques enhance efficiency and geometric adaptability; for dispersion particle-fueled reactors, improved global-local methods can handle double heterogeneity; and for fast reactors, methods like TONE and point-wise/ultra-fine-group approaches effectively manage the hard spectrum and multiple resonant nuclides.For solving the neutron transport equation, NECP Lab has advanced both deterministic and stochastic (Monte Carlo) numerical methods. Developments on deterministic methods include stabilized finite element methods and transmission probability methods on triangular meshes for complex geometries, as well as enhanced method of characteristics (MOC) schemes with improved ray-tracing, CPU-GPU heterogeneous parallelism, and 2D/1D coupling for 3D simulations. For Monte Carlo methods, techniques like consistent adjoint driven importance sampling (CADIS) and its variants, along with adaptive weight window generation, drastically improve the computational efficiency for deep-penetration radiation shielding problems.Based on these theoretical advancements, the NECP Lab has independently developed a full-fledged suite of high-fidelity neutronics analysis software with complete intellectual property rights. Key software packages include: NECP-Atlas: for processing evaluated nuclear data files (including the modern GNDS format) into application-ready libraries. NECP-Bamboo Series: for PWR core physics analysis, including the industrially validated Bamboo-C and the high-resolution pin-by-pin code NECP-Bamboo2.0. NECP-X: a high-fidelity deterministic numerical reactor code capable of full-core steady-state, transient, and depletion simulations with multi-physics coupling. NECP-SARAX: a fast reactor analysis code system already validated against international benchmarks and zero-power/reactor start-up tests. NECP-Panda: for the physics analysis of pebble-bed HTGRs, incorporating treatments for double heterogeneity. NECP-Hydra & NECP-MCX: for radiation shielding calculations using deterministic (discrete ordinates) and hybrid Monte-Carlo/deterministic methods, respectively, both recognized as fully independent and controllable industrial software.These software systems have been extensively applied to support the design and operation of various advanced reactors in China. Applications include the nuclear design of special micro-reactors, ex-core neutron field analysis for large pool-type sodium-cooled fast reactors, and several operational support technologies for PWR nuclear power plants. The latter includes dynamic flux mapping, single-point calibration of ex-core detectors, and improved dynamic rod worth measurement techniques, which have enhanced both the safety and economic performance of operating reactors.