Modelling of reflood phase of a loss of coolant accident with COBRA-TF code and analysis of model uncertainities

Best estimate codes used in the safety analysis, licensing, research and development fields frequently are cost efficient, quick and reliable tools. For these codes which are widely used by the industry, universities and public intitutions, it is very important to validate the physical models against experimantal data which are obtained for wide range of conditions. Therefore, the modifications which improve solution methods, increase the accuracy in code predictions and performance are required by developing new models and adapting them to the codes. In this dissertation, reflood phase of postulated Loss of Coolant Accident in Pressurized Water Reactors is analyzed and modifications were performed to analyze the maximum clad temperature behaviour, one of the most important safety criterion. After specifiying weaknesses of reflood models in COBRA-TF and underlying compenents causing them and applying corrective and improving models to the code it is aimed to obtain better results with respect to FLECHT-SEASET (Full Length Emergency Core Heat Transfer-System Effects and Separate Effects Tests) and RBHT (Rod Bundle Heat Transfer) experimental data. In the study, it is observed that COBRA-TF calculates the Minimum Film Boiling Temperature (T_min) inconsistently when it's compared with the experimantal data. Since T_min model affects the heat transfer behaviour completely, applying a more advanced T_min model, better code results and consistent fuel temperature behaviour were obtained. As the results of the analysis carried out after this modification, it is observed that the film entrainment in annular film flow and quench front entrainment generated on cooling fluid surface models existing in COBRA-TF have weaknesses on estimating some experimental conditions. Therefore, adding different models to available entrainment calculation had been applied to COBRA-TF code, additionally. The results of using new models, FLECHT SEASET and RBHT experimental data were compared, and significantly modified results were obtained in calculating quench front, fuel clad temperature and vapor temperature.