An IFDF accelerated parallel nodal SN method for XYZ geometry in SARAX code system

•Simplified high-order transverse leakage terms for XYZ nodal SN scheme.•IFDF acceleration method with adaptive diffusion coefficients.•Spatial parallel algorithm with domain decomposition and MPI communications.•Encouraging results for two fast reactors with square assemblies. There are some fast r...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Annals of nuclear energy 2022-02, Vol.166, p.108710, Article 108710
Hauptverfasser: Xu, Zhitao, Zheng, Youqi, Wu, Hongchun
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:•Simplified high-order transverse leakage terms for XYZ nodal SN scheme.•IFDF acceleration method with adaptive diffusion coefficients.•Spatial parallel algorithm with domain decomposition and MPI communications.•Encouraging results for two fast reactors with square assemblies. There are some fast reactors designed with square assemblies requiring whole-core multigroup neutron transport solutions. A simplified method to calculate the high-order transverse leakage terms was proposed based on the NEFD nodal SN scheme in XYZ geometry. The IFDF acceleration method with adaptive diffusion coefficients and parallel algorithm with spatial domain decomposition and MPI communications were employed to solve the discretization system efficiently. A small fast reactor and the ZPPR-10B large fast reactor were calculated. Encouraging accuracy improvements were obtained for both the eigenvalue and the distribution quantities with only about 10% additional calculation cost for the simplified high-order transverse leakage terms. The 4-group small fast reactor was solved in about 2 s with 16 CPU cores and the 33-group ZPPR-10B large fast reactor was solved in about 19 s with 196 CPU cores. Speedup beyond 60 was obtained compared with the traditional serial fission-source iteration on a 16-CPU workstation.
ISSN:0306-4549
1873-2100
DOI:10.1016/j.anucene.2021.108710