• ATWS-I
• LOCA
• Quenching
• Rewet
• subchannel
• thermo-hydraulics

During normal operation in a Boiling Water Reactor (BWR), a constant contact between a liquid film and the fuel rod surface is maintained thus guaranteeing efficient cooling of the cladding surface. During a Loss-Of-Coolant Accident (LOCA) or an Anticipated Transient Without Scram-Instability (ATWS-I), the established liquid film on the fuel rods can be vaporized due to coolant inventory loss. This inventory loss can be local (ATWS-I) or at a larger scale (LOCA). When the film is vaporized, optimal heat transfer at the cladding surface is lost and fuel bundles begin heating-up rapidly. To avoid high surface temperatures that could compromise the integrity of the cladding, the film of the coolant needs to be re-established on the surface. This phenomenon is called rewetting and its modeling is very important for determining the thermal behavior of the cladding. In order to predict the cladding temperature behavior and the rewetting of the fuel rods, Nuclear Safety Analysis (NSA) calculations need to be performed.
Traditionally, for Appendix K compliant methods [14], LOCA transients have used conservative thermal-hydraulic models. The simplified computer models and often excessive conservatism of these analytical methods do not allow to fully capture the details of the physical phenomena during a LOCA transient. In addition to unrealistic modeling, this conservatism leads to unnecessary large thermal margins that may impact plant economics and operation. While ATWS-I is a beyond design basis event and best-estimate conditions can be used for its analysis, both LOCA and ATWS-I rely on the numerical prediction of the rewetting. Therefore, accurately predicting flow regimes and heat transfer regimes is critical to the safety analysis of nuclear reactors and rewetting in particular. The predictions of the rewetting phenomenon requires a variety of constitutive relationships to describe the mass, momentum, and energy exchange that occurs between the flow fields (steam, droplets and liquid film) and provide closure to the set of momentum equations. Rewetting is characterized as a localized physical phenomenon where sharp temperature, void fraction and mass flux gradients are observed within a small distance. Often, these characteristics are not well represented by Appendix K compliant methods or overly conservative analyses in general.
The optimal utilization of modern high performance fuel designs in combination with the industrial
application of best-estimate methodologies generated interest to depart from these conservative
approaches. However, most thermal-hydraulic system codes used throughout the nuclear industry for
accident analysis employ a one-dimensional two-field two-phase model that tracks single homogeneous
liquid and vapor phases and relies on simplified models or correlations to predict the complex heat
transfer and flow phenomena during quenching. Several best-estimate models used in NSA codes have a
fine mesh option for the heat conduction solution (two-phase two-field codes such as TRACE,
RELAP and two-phase three-field subchannel codes such as COBRA)but the very small
mesh sizes required to characterize the temperature gradients at the quench fronts makes such method
impractical for system calculations. Consequently, the multiscale characteristics of the quench front could be lost and important aspects of the physics of this phenomenon may not be captured. Improving the predictive capabilities of ECCS models is crucial for safety analysis since ECCS capacity may dictate core operating limits on local power production.
The originality of the present study lies in that it has examined the predictive capabilities of two-field and three-field transient analysis codes focusing on the rewetting region for BWR at LOCA and ATWS-I conditions. This analysis of the rewetting characteristics yielded the development of a quenching model that includes:
1) A dynamic axial re-nodalization scheme of the local quench front. A coupled T-H re-nodalization
model with a fine moving mesh dynamically refining the Eulerian hydraulic mesh of an NSA
computer code was developed to better model quenching heat transfer. This refined Eulerian
hydraulic mesh is in turn coupled with a Lagrangian non-uniform nodalization in order to
compute the temperature distribution within the fuel rod and calculate the heat transfer to the
coolant. The Lagrangian-Eulerian coupling and dynamic re-nodalization logic for three-field
solvers is a new approach that has been developed for this thesis.
2) The development of a heat transfer logic and the use of pr-established heat transfer correlations to properly calculate temperature gradients at locations where sharp thermal-hydraulic property
changes are experienced,
3) The development of a 2D-Conduction Controlled model that solves the temperature distribution
in the fuel rod and computes film front velocities, heat transfer from the rod surface to the fluid
and couples these results to the considered NSA computer code original T/H solution.
The three-field GEH proprietary code COBRAG code was used for this work because of its capability to
model steam, droplets and film which provides additional resolution in modeling the rewetting
phenomenon. Its spatial resolution capabilities were increased with including the dynamic re-nodalization scheme and the implementation of the quenching model. In order to incorporate the models discussed previously, COBRAG was modified to include constitutive relations to compute the flow regimes and heat transfer, mass and energy distribution between fields at the rewetting front.. The models that are incorporated into the proposed three-field rewetting package are developed uniquely for the problem at hand.
The developed rewetting modeling package for COBRAG includes:
1) The implementation of a functional relationship modeling the liquid phase split between film and
droplet based on comparisons to experimental data. This model was implemented in COBRAG to
refine the liquid distribution between the liquid film and the droplets.
2) The improvements to the numerical method to mitigate water packing in numerical cells at
locations of sharp interfacial property changes and low pressure applications.
The quenching model and the methods package developed during the research work provide a detailed
spatial representation of the subchannel thermal-hydraulic properties for a single BWR bundle for
ATWS-I and LOCA conditions. Given the importance of the physical phenomena at play in a BWR
channel for safety analyses of various reactor accident scenarios and the interest to generate more precise and updated experimental data, this modeling methodology provides a starting ground in improving the predictive capabilities of three-field transient analysis codes until a more stable and viable approach is ascertained for both ATWS-I and LOCA applications. Benchmark results in the current study demonstrate that the development and inclusion of this newly proposed model for rewetting modeling along with the necessary COBRAG modifications is able to match experimental data for high pressure and high temperature quenching. The development of the 2D-Conduction Controlled quenching model and renodalization scheme along with the COBRAG modification results in an enhanced capability to model the rewetting phenomenon in a Lagrangian-Eulerian framework for two-phase flow and three fields numerical model.
The result of this thesis developed an advanced COBRAG code version with transient capabilities. It was demonstrated that this advanced COBRAG code reasonably models experimental results representative of BWR LOCA and ATWS-I analyses.