Development of a hybrid method to improve the sensitivity and uncertainty analysis for homogenized few-group cross sections

Research output: Contribution to journalArticlepeer-review

2 Scopus citations

Abstract

In the framework of two-step method of reactor core calculation, few-group homogenized cross sections generated by lattice-physics calculations are key input parameters for the three-dimensional full-core calculation. Conventional method for few-group cross-sections sensitivity and uncertainty (S&U) analysis related to the nuclear data was performed based on the effective self-shielding cross sections instead of the continuous-energy cross sections, which means resonance self-shielding effect (implicit effect) is neglected. Furthermore, the multi-group covariance data is generated from the continuous-energy cross sections. Therefore, in order to perform S&U analysis with respect to the continuous-energy cross sections for both accuracy and consistency, a hybrid method is proposed in this paper. The subgroup-parameter sensitivity-coefficients are calculated based on the direct perturbation (DP) method. The sensitivity-coefficients of the effective self-shielding cross sections and the responses (k eff and few-group homogenized cross sections) are calculated based on the generalized perturbation theory (GPT). A boiling water reactor (BWR) pin-cell problem under different power conditions is calculated and analyzed. The numerical results reveal that the proposed hybrid method improves the sensitivity-coefficients of eigenvalue and few-group homogenized cross sections. The temperature effects on the sensitivity-coefficients are demonstrated and the uncertainties are analyzed.

Original languageEnglish
Pages (from-to)769-783
Number of pages15
JournalJournal of Nuclear Science and Technology
Volume54
Issue number7
DOIs
StatePublished - 3 Jul 2017

Keywords

  • Nuclear data
  • generalized perturbation theory
  • sensitivity and uncertainty
  • subgroup resonance calculation method

Fingerprint

Dive into the research topics of 'Development of a hybrid method to improve the sensitivity and uncertainty analysis for homogenized few-group cross sections'. Together they form a unique fingerprint.

Cite this