Title:
On-the-fly Generation of Spatial Transfer Functions for Efficient Monte Carlo-TH Coupled Calculations

Thumbnail Image
Author(s)
Terlizzi, Stefano
Authors
Advisor(s)
Kotlyar, Dan
Advisor(s)
Person
Editor(s)
Associated Organization(s)
Series
Supplementary to
Abstract
A novel methodology to efficiently perform coupled Monte Carlo (MC)- Thermal Hydraulics (TH) is developed in this thesis. This goal is achieved by accelerating the classical Picard iteration scheme using a deterministic, perturbation-based prediction step. The latter produces an improved initial guess of the power and the corresponding TH fields, i.e., temperature and density, at the next MC calculation, therefore accelerating the convergence of the whole algorithm. The prediction relies on a novel technique for the calculation of the cross-sections’ spatial distribution due to variations in the TH scalar fields. The latter technique is an original and novel contribution that relies on the definition of generalized transfer functions (GTF) to calculate the cross-sections given a change in density and temperature profiles. The method has many desirable characteristics. First, it allows to capture non-local effects contrarily to current methods for cross-sections interpolation/reconstruction. Second, the GTF, despite being here applied to MC codes, is general and may be used for deterministic methods as well, e.g., GRIFFIN. Finally, the method is geometrically agnostic. Thus, it can be utilized to study non-traditional geometries, i.e., not square or hexagonal, that characterize many advanced nuclear reactors’ concepts. When applied to perform coupled MC-TH analysis on mono-dimensional and three-dimensional Light Water Reactors (LWR)-based computational problems, a speedup factor between 2.1 and 12.4 was observed with respect to the standard Picard iteration scheme.
Sponsor
Date Issued
2020-12-06
Extent
Resource Type
Text
Resource Subtype
Dissertation
Rights Statement
Rights URI