Some of the most important challenges in the field of Monte-Carlo (MC) reactor analysis are the integration of thermal-hydraulic feedback, the extension of the method to the study of transient phenomena, and the convergence acceleration of the MC algorithm for the analysis of reactor criticality. This thesis tries to meet these three challenges by suggesting algorithms that can cope with the encountered issues.
As a first step, this work investigates the insertion of Thermal-Hydraulic (T-H) feedback to static Monte-Carlo. Initially, a “serial” coupling scheme, that corresponds to the sequential execution of the involved solvers, is developed to provide reference results. Then, this work suggests the use of an approximate Newton coupling algorithm. The motivation for this approach is the interest in an algorithm that can maintain the distinct treatment of the involved fields of physics within a tight coupling context. This work investigates the behaviour of the proposed method when the open-source MC neutronics code OpenMC is coupled with the T-H code COBRA-EN. The performance and the accuracy of the proposed coupling scheme are evaluated and compared with those of the traditional serial iterative scheme. The results show a significant numerical improvement leading to more accurate results.
Secondly, this thesis investigates the development of a Monte-Carlo dynamic module in OpenMC for the analysis of transient phenomena. A straightforward physical treatment of a transient problem requires the assessment of the temporal evolution of the simulated neutrons, which is no present in static Monte-Carlo; however, this is not adequate. To properly analyze transient phenomena, the simulation of delayed neutrons and other necessary extensions and modifications are needed. The selected method has been recently proposed in the literature and is here inserted in OpenMC following the code’s features. Hence, an extra challenge that this work meets is the desire for an optimum embodiment in OpenMC, minimizing the necessary modifications and maximizing the advantage resulting from its existing capabilities. Moreover, the addition of dynamic T-H feedback is investigated. The key points of the developed module, as well as the results of the analysis of various numerical experiments, are presented and discussed. The results confirm the successful development of the dynamic Monte-Carlo module, pointing out its capability to effectively analyze various reactor core transients.
Finally, a new convergence acceleration method of the Monte-Carlo classical Source Iteration (SI) is presented. Whereas the classical SI guarantees the convergence to the fundamental eigenmode, very often the convergence is slow. In this thesis, an alternative version of the traditional Monte-Carlo SI algorithm is formulated, developed, and analyzed to accelerate the Monte-Carlo criticality analysis numerically. More specifically, the Jacobian-Free Newton Krylov method is adopted in the Monte-Carlo k-eigenvalue context to accelerate the convergence. The method is evaluated in three test cases showing better performance than the traditional Coarse-Mesh Finite-Difference acceleration technique.