A perturbation-based acceleration for Monte Carlo – Thermal Hydraulics Picard iterations. Part II: Application to 3D PWR-based problems

Abstract

This paper focuses on the application and demonstration of the GTF-based acceleration introduced in the first part of the article to three-dimensional problems. Two novelties are here introduced over the state-of-the-art. First, an operative definition of generalized transfer functions for three-dimensional distributions is introduced. Second, the GTF-based acceleration is used to speed up the Picard iteration scheme applied to 3D problems based on Pressurized Water Reactor (PWR) designs. Consistently with the mono-dimensional results presented in Part I of this article, the GTF-based acceleration allows to achieve convergence within a lower number of iterations, with a speed-up factor up to 5.5, compared to the original Picard iteration.