Effect of porous media on heat removal capability during normal and severe accident conditions of PWR

Abstract

Various porous media in pressurized water reactors (PWR) can affect the heat removal capability of nuclear fuel. In the first part of this study, the models for the CRUD that is formed on the fuel surface during the normal operating condition of PWR was suggested. The model was proposed based on the phenomenon that the flow paths of the liquid and the gas are separated due to the structural characteristic of the CRUD. First, it was shown that the CRUD have heat transfer regimes such as liquid-saturated CRUD, wick boiling, and film boiling through the classification of the WALT experimental data [6]. In the wick-boiling model, it is assumed based on the visual observations that the heat is removed through nucleate boiling near the fuel rod surface, while most of the previous models [1-5] assumed evaporation from the wall of the steam chimney. As a result, the prediction ability for the heat transfer was improved from 22% root-mean-square error (RMSE) of the previous model [1] to 12% RMSE. In the film-boiling model, it is assumed that the capillary pressure changes with applied heat flux based on the previous observation, unlike the assumption of constant capillary pressure of the previous model [7]. As a result, the prediction on the wall superheat from the previous model showed the maximum error of 15°C compared to the WALT experiment [6], while one from the suggested film-boiling model showed errors within about 3°C. In addition, we also suggested the model for the onset of transition to the film-boiling regime. In the suggested model, it is assumed that the transition begins when the void fraction in the thermal boundary layer reaches the critical value, which is found to be 0.55. As a result, the heat flux for the onset of the transition was able to be predicted within 18% RMSE comparing to the existing experiment [6]. Now, through the models suggested for the CRUD, the heat removal capability of the CRUD can be predicted for all heat transfer regimes found in the normal operating condition of PWR. In addition, analysis on the issue of CRUD Induced Power Shift (CIPS) was able to be conducted coupling the current models to the transport equations for the chemicals in the coolant. In the second part of this study, the similar model proposed in the first part was applied to quenching analysis in the debris bed and the crust of the molten corium during severe accidents. Unlike the existing models based on the assumption that the quenching rate is limited due to the counter current flow limit (CCFL) phenomenon [8, 9], the current model is based on three different main features: (i) the separation of the fluid paths, (ii) the hydraulic limit of quenching rate caused by the limitation in the capillary and the buoyancy forces, and (iii) the thermal limit of quenching rate caused by the high temperature of the debris bed. As a result, the current model well predicted the decreasing trend of heat flux with increasing bed temperature. Also, the RMSE of the current model was 15%, which was greatly improved from 38% of the existing model [8, 9]. In order to develop the quenching model for the fractured crust, the morphology of the fractured crust first was determined by two short-term thermal-stress based criteria: a tensile-strength criterion and a toughness criterion for the fracture while all the previous models utilized the long-term creep criterion developed by Epstein [10, 11]. When the current model was used to estimate the heat removal rate during the quenching of the fractured crust, it showed 26% of RMSE for data [11], which was improved from 39% RMSE of the existing model [11] based on the CCFL concept to estimate the quenching rate.