32 Müller, Henrik Allelein, Hans-Josef Kneer, Reinhold Laurien, Eckart 413110 RWTH Aachen University Aachen English 1 Online-Ressource : Illustrationen, Diagramme The use of numerical methods for the safety analysis of nuclear power plants helps suppliers, operators and safety authorities to justify the plant layout, understand the plant behavior in different accident scenarios, to define accident mitigation measures and to identify potentials for the improvement of the plant safety. Complementary to the well-established lumped parameter or system codes, there is an increasing demand to use computational fluid dynamics (CFD) codes for this task. CFD codes offer the ability to gain additional insights into, for example, localized or three-dimensional flow phenomena in the course of an accident, especially the distribution of hydrogen and the formation of combustible mixtures. One application of great interest for the use of CFD-codes is the containment flow during a loss of coolant accident (LOCA). Due to the presence of phenomena such as wall condensation, the analysis of the containment flow during a LOCA requires a fine mesh resolution at the wall surface to accurately predict the flow. However, with such a fine mesh, the calculations require a prohibitive large amount of computation time. Therefore, subgrid models, so called wall functions, need to be employed to reduce the required grid resolution and thus the computational time. Currently available wall functions are not suitable for modeling containment flows under LOCA conditions. In particular, they do not include the effects of buoyancy and condensation-induced wall normal velocity, commonly referred to as the suction effect. The wall function approach presented in this thesis consists of a set of empirical functions based on experimental data supplemented by high resolution numerical data. In addition to the effects typically encountered in wall functions, the approach includes the influence of buoyancy and wall normal flow. To account for these effects, a covering set of non-dimensional parameters is derived and the input data are combined into multivariate algebraic functions using the concept of radial basis functions. This allows an easy and computationally fast integration of the wall function into different CFD codes. An exemplary implementation in the CFD code ANSYS CFX is presented. A first validation is performed based on separate and integral effect tests. The separate effect test validation demonstrates that the newly developed wall function approach provides improved results compared to a standard wall function, especially under containment-like flow conditions. However, larger scale integral effect tests do not show such improvement due to the limited application range of the new model resulting from the limited amount of available input data. Nevertheless, the results show that the newly developed wall function approach has the potential to enable computationally affordable CFD calculations of a full containment under LOCA conditions with an accurate prediction of the near wall momentum, heat and mass transfer, although some work remains to be done. Veröffentlicht auf dem Publikationsserver der RWTH Aachen University 2023 ; Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2022 ; 2, ; PUB:(DE-HGF)11, ; Dissertation / PhD Thesis Dissertation Rheinisch-Westfälische Technische Hochschule Aachen 2022 2022 RWTH-2023-06349 2022 https://publications.rwth-aachen.de/record/960707