Heterogeneous Parallel High-order Scattering MOC and Its Application to Simulation of Critical Experiment

Abstract

The purpose of this study is to investigate the impact of neutron anisotropic scattering on critical experimental setups and to develop a MOC (method of characteristic) program capable of handling anisotropic scattering, along with a high-performance heterogeneous parallel algorithm for high-order scattering transport calculations. In the initial stages, the physical calculations of the critical experimental setup were analyzed, revealing that neutron anisotropic scattering can significantly affect the calculation results, particularly when a thicker water reflector is present. Building upon the P1 anisotropic scattering MOC, a specialized MOC program was developed to address this issue. To validate the accuracy of the newly developed program for critical experimental simulations, the researchers selected the LCT011 critical experimental benchmark for neutronic calculations. A comprehensive comparison was performed between the results obtained from the MOC program and a Monte Carlo program, serving as a benchmark for verification. One notable challenge encountered during the study was the substantial increase in computation time and memory consumption caused by the presence of anisotropic sources. This created a significant memory burden, especially on heterogeneous systems. Consequently, the researchers conducted a thorough performance analysis of the high-order scattering transport solver employed in the program. The numerical results obtained from the study showcase that the MOC program achieves comparable accuracy to the Monte Carlo program under conditions involving high-order scattering computations. Furthermore, the researchers observed that the developed program exhibited remarkable computational efficiency, making it a promising alternative to the Monte Carlo method. By effectively addressing the impact of neutron anisotropic scattering and providing accurate results with enhanced computational efficiency, the developed MOC program holds great potential for advancing critical experimental simulations. This research significantly contributes to the field of physical calculations by offering a reliable and efficient solution for handling anisotropic scattering in high-order transport calculations. In conclusion, this study presents the purposeful investigation of neutron anisotropic scattering in critical experimental setups, resulting in the development of a specialized MOC program and a high-performance heterogeneous parallel algorithm. The validation process, conducted using the LCT011 critical experimental benchmark, confirms the accuracy of the program. The performance analysis showcases the computational efficiency of the developed program, thus establishing its viability for critical experimental simulations involving anisotropic scattering effects. This research underscores the importance of accurate neutron anisotropic scattering calculations and offers an innovative solution to address the associated challenges in the field of reactor core physical calculations.

Keywords

• method of characteristic
• high-order scattering
• critical experiment
• heterogenous parallel