The potential consequences of hydrogen release and combustion during a severe accident in a Light Water Reactor (LWR) have received considerable attention after the Fukushima accident. The pressure loads resulting from hydrogen combustion can be detrimental to the structural integrity of the nuclear reactor safety systems and the reactor containment. Therefore, accurate prediction of these pressure loads is important from a safety point of view. The considered pressure loads are determined by the turbulent flame acceleration. This flame acceleration is determined by the amount of turbulence generated locally during the combustion process. Lumped parameter codes can only assume, but cannot compute this local turbulence generation process due to their inherent limitations. Therefore, a three dimensional CFD based approach is needed. In this paper, it is demonstrated that it is of utmost importance to apply successive mesh and time step refinement systematically in CFD analyses for the considered application. This has been applied in this paper. Within the field of hydrogen safety, we have not found CFD analyses published earlier in the open literature that were based on successive mesh and time step refinement. In the light of the above, the lack of the demonstration that the numerical errors are negligible in the CFD analyses of hydrogen deflagrations must be considered as a serious shortcoming. Therefore, the authors are developing a CFD methodology that can be used for the analyses of hydrogen deflagrations in experimental facilities as well as in full scale reactor containments, with which it can be demonstrated efficiently that the mesh and time step requirements are fulfilled. Such a CFD methodology is deemed to be essential in order to be able to prove containment integrity under severe accident conditions. The authors are well underway with the development of this methodology, and the first part of the development will be presented in this paper.
Furthermore, it is demonstrated that the turbulent flame acceleration, and therefore the resulting pressure loads, depend also strongly on the initial turbulence present in the ignition region. Therefore, it is recommended that for future validations, hydrogen deflagration experiments will be performed where this initial turbulence is accurately measured. (C) 2012 Elsevier B.V. All rights reserved.