Answers are grounded in this paper and cite their sources.
The AP1000 reactor utilizes a complex control rod (CR) strategy for both reactivity control and daily load-following operations, characterized by frequent and incremental CR adjustments. These operational requirements pose significant challenges to achieving both high fidelity and computational efficiency in core simulations. To address these challenges, the neutron transport code FLASH, which utilizes the fission response function method, was enhanced in this study to enable high-fidelity AP1000 simulations under various CR insertion conditions.Specifically, an adaptive core slicing algorithm was developed to automatically partition the core into axial layers based on actual CR configurations. In addition, a local submesh cusping treatment was implemented to represent CR tips located within axial computational meshes.To further improve computational efficiency in complex scenarios, the SLEPc high-performance numerical library was integrated to solve the large-scale sparse eigenvalue problem using the Krylov-Schur eigensolver. Verification against Monte Carlo reference solutions for four typical AP1000 CR insertion scenarios demonstrated that the keff deviations remained within 260 pcm and the pinwise fission rate root-mean-square error values remained below 1.2%.An additional comparison showed that for the adopted 1-in. axial mesh, the maximum difference in the keff between the calculations with and without cusping treatment did not exceed 9 pcm, while the treatment provided a more detailed representation of local fission rates near the rod tips by resolving the rodded and unrodded portions of partially rodded meshes.Furthermore, the SLEPc implementation reduced the execution time by approximately 20% for the most complex 15-slice scenario, which was simulated in approximately 2.5 min. These developments demonstrated that FLASH provides a robust platform for high-fidelity, quasi-real-time analyses of the AP1000 core.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: