Erosion of a stably stratified layer by a turbulent round jet

In the course of a severe accident in the nuclear power plant, hydrogen can be produced by a chemical reaction between steam and Zircaloy cladding. The hydrogen gas can escape into the reactor containment, either through a leak in the primary cooling circuit or due to an intended release, in order to prevent over-pressure in the reactor pressure vessel. Due to the different densities of air and hydrogen, stable stratification is expected, hydrogen above air, in the upper part of the containment. Subsequent steam-hydrogen mixture release, in the form of a vertical turbulent jet, will erode the stably stratified layer in the top part of the containment. In the course of that erosion process, ignition limits of hydrogen-air-steam mixture can be reached. In March 2011 steam-hydrogen mixture release has led to the explosions in the Fukushima Daiichi boiling water reactors during the tsunami accident. Knowledge of hydrogen distribution inside the containment is needed in order to design and optimize the positioning of countermeasures, like recombiners. According to Karwat et al. it is not practical to convey sever accident experiments in full scale reactors. Thus, scaled-down facilities were built in order to mimic the break-up of a stable stratified layer by a turbulent jet, e.g. PANDA vessel in PSI, Switzerland and MISTRA, CEA-Saclay, France.The internal dimensions of the MISTRA facility, 4.25 meters in diameter and 7.4 meters in height, correspond to a length scale ratio of 0.1 compared to a typical French PWR. Though reduced in dimensions, both above-mentioned systems are considered as rather large-scale experimental facilities. In the experiments performed in MISTRA and PANDA, hydrogen is replaced by helium, due to the safety reasons, and its concentration is monitored during the evolution of the experiments. In addition to the valuable physical insights that those facilities provide, they serve as validation experiments for CFD simulations. The capability of CFD tools to predict nuclear power plant containment flows was assessed in a number of validation benchmarks.The LOWMA-3 experiment was performed in the MISTRA facility. In that experiment, helium was injected slowly from the upper part of the facility and a helium-rich layer was set. When the evolved velocities were settled down, air jet was issued from a 72 mm vertical tube towards the helium layer from below. The Reynolds number based on the tube diameter was 14,000, the interaction Froude number was of order of 1 and the distance from the air injection level to the helium layer was about 20 tube diameters. The air jet injection phase lasted for 6,000 seconds. During the experiment, the helium concentration evolution was measured at 10 measurement points (denoted tcg20-tcg29), off-center with respect to the jet axis. This work examines numerically the concentration of helium in LOWMA-3 experiment using the commercial CFD code, ANSYS Fluent. The numerical approach includes a separate effect study of a free turbulent round jet in order to validate the velocity decay and the spreading rate of the predicted jet flow. As a preliminary study, modeling of a free turbulent jet is addressed and compared with well-known experiments from the literature. The following step is modeling of the first 1000 seconds into the transient of the helium erosion phase, as this period of time is assumed to be the most significant in this experimental validation of the CFD results. It is believed that if the transient behavior of the results is captured during this time period, the rest of the transient should also be predicted correctly. The separate effect study as well as the LOWMA-3 simulations results and conclusions are addressed in the next sections