CFD code-to-code benchmarking on simulation of fire and smoke propagation for nuclear applications

Abstract

Abstract The nuclear industry has seen an increased use of Computational Fluid Dynamics (CFD) technology as a high-fidelity tool for design-basis and beyond design-basis accident simulations. For fire hazards, detailed experimental investigation in a realistic environment is often impractical. Therefore, the use of an advanced 3D CFD modeling approach to predict smoke and fire propagation has risen, especially for nuclear power plant fire modeling applications. CFD simulations can include major contributing and mitigating mechanisms regarding the fire propagation and its consequences, in fine detail and under realistic geometry conditions. In addition, accurate simulations of unsteady 3D fire and smoke spread, in conjunction with evacuation analyses, can be used to establish effective emergency evacuation strategies. In the present study, the fire and smoke propagation in a main control room was modelled using Reynolds-Averaged Navier-Stokes Equation based twoequation turbulence models and a Large Eddy Simulation model implemented in a commercial CFD code, Siemens STAR-CCM+. The predictions from STAR-CCM+ were compared with those obtained in a previous study using the NIST Fire Dynamic Simulator (FDS) code. Although higher temperature was predicted at the fire source by STAR-CCM+, the current study predicted trends similar to the FDS results with some minor differences at the operator location. A follow-up study is currently underway to investigate the factors affecting the differences, including further assessment of codes against available benchmark experiments.