A Shared-Memory-Based Coupling Scheme for Modeling the Behavior of a Nuclear Power Plant Core

Abstract

The different phenomena that take place in the core of a nuclear reactor comprise a wide variety of physical aspects, which differ both in their nature as in their complexity as well. In general, these effects can be divided into four main disciplines, namely neutronics, thermal-hydraulics, plant conditions and control. Even though neutrons interact with matter in a pretty well-known way, the equations that model how they behave in a nuclear reactor derived from basic mechanistic considerations cannot be yet—and probably ever–be solved completely. And, even more, first-principle equations for turbulent two-phase flow still are not even available. Therefore, to design and analyze nuclear reactors some kind of simplifications ought to be used, which—despite its usual meaning—may still be challenging from a computational point of view. Being the equations behind these four disciplines radically different in their mathematical characteristics, there usually exist dedicated codes that solve each problem separately, especially for the most-complex tail of the spectrum of simplifications. It is therefore desired to devise a mechanism for coupling these particular codes in such a way that the different dependencies are explicitly addressed. A shared-memory-based coupling mechanism was developed in order to perform the aforementioned tasks, which is currently being used in the nuclear industry although it may have applications in other fields as well.