Validation and application of the MEWA code to analysis of debris bed coolability

Abstract

This paper was first aimed at validating the MEWA code against experiments for two-phase flow and dryout in particulate beds, and then investigating the coolability of ex-vessel debris beds with cylindrical, conical and truncated conical shapes assumed to form under severe accident scenarios of a boiling water reactor. The validation was mainly performed against the POMECO-FL and POMECO-HT experiments carried out at KTH for investigating frictional laws and coolability limit (dryout) of particulate beds, respectively. The comparison of the experimental and numerical results shows that the MEWA code is capable of predicting both the pressure drop of two-phase flow through porous media and the dryout condition of various stratified beds. While the coolability of a one-dimensional homogeneous debris bed is bounded by counter-current flow limit (CCFL), the coolability of a heap-like debris bed can be improved due to lateral ingression of coolant in a multi-dimensional geometry. The simulations showed that the dryout power density of a prototypical debris bed was roughly inversely proportional to the bed’s height regardless of the bed’s shape. The impacts of a debris bed’s features on coolability are manifested in three aspects: multidimensionality and contour surface area of the bed, as well as the uniformity of its shape. The contour surface area is defined as the interface between debris bed and water pool, and its effective value depends on the surface orientation that determines the amount of water ingress and vapor escape. The perfect uniformity in bed’s shape as cylindrical bed results in even distributions of temperature and void fraction. The dryout power density was also predicted to be strongly correlated to the uniformity of bed’s shape. The MEWA simulation also predicted that coolability was improved by an downcomer embedded in the center of debris bed. The efficiency of such enhancement was largely determined by the downcomer’s length, whose optimal value was obtained in simulation.