Nuclear reactor severe accidents can lead to the release of a large amount of radioactive material and cause immense disaster to the environment. Since the Fukushima nuclear accident in Japan, the severe accident research has drawn worldwide attention. Based on the one-dimensional heat conduction model, a DEBRIS-HT program for analyzing the heat transfer characteristics of a debris bed after a severe accident of a sodium-cooled fast reactor was developed. The basic idea of the DEBRIS-HT program is to simplify the complex energy transfer process in the debris bed to a simple one-dimensional heat transfer problem by solving the equivalent thermal conductivity in different situations. In this paper, the DEBRIS-HT program code is prepared by using the existing model and compared with the experimental results. The results show that the DEBRIS-HT program can correctly predict the heat transfer process in the fragment bed. In addition, the heat transfer characteristics analysis program is also used to model the core catcher of the China fast reactor. Firstly, the dryout heat flux when all of molten core dropped on the core catcher was calculated, which was compared with the result of Lipinski’s zero dimensional model, and the error between two values is only 11.2%. Then, the temperature distribution was calculated with the heat power of 15MW.

The classic four-equation drift-flux model treats the two-phase flow as a mixture to formulate mass, momentum, and energy balance equations. The dispersed phase is modeled in the mass balance equation to describe the mass transfer between the two phases. However, this model implies an assumption of the thermal equilibrium in the mixture energy balance equation. Therefore, a five-equation drift-flux model and its constitutive equations have been developed to relax the thermal equilibrium assumption in the classic four-equation drift-flux model when it is applied to thermal non-equilibrium phenomena in Light Water Reactors (LWRs). The additional energy balance equation of the dispersed phase has been introduced to address thermal non-equilibrium phenomena in LWRs in the five-equation drift-flux model. In addition, the newly added energy equation for the dispersed phase does not change the basic kinematic parameters and uses the mixture velocity instead of the dispersed phase velocity. The hyperbolicity of the drift-flux model system equations is investigated by examining if the partial differential equations have only real eigenvalues. The results show that the five-equation drift-flux model is a well-posed hyperbolic model, which can be solved by both advanced solvers using the finite element method, such as MOOSE Framework, and the traditional solvers using finite difference method, such as RELAP5 code. In developing the thermal non-equilibrium five-equation drift-flux model, it is necessary to select, improve, or develop constitutive models as closure relations for variables used in the field equations. In our current study, a literature survey has been conducted to review appropriate constitutive models and correlations for the dispersed phase mass generation rate, interfacial energy transfer, distribution parameter, and drift velocity in the five-equation drift-flux model.

The purpose of this study is to propose a new formula for evaluating the local damage to reinforced concrete structures caused by oblique impact based on past experimental results and simulation results. In this paper, we present the results of simulation analyses for evaluating the perforation of concrete panels due to oblique impact by deformable projectiles. Various response characteristics and perforation mechanisms such as detailed perforation behavior, damage to reinforced concrete panels, rupture states of the deformable projectiles, reduction in projectile velocity, and residual velocity of the projectile after perforation, and energy transfer processes are clarified. Especially, it is found that the sliding energy consumed by friction occurring at the contact surface between a projectile and a reinforced concrete panel due to oblique impact is remarkably larger than that due to normal impact.

The flow-induced vibration of a pipe is an important issue in various engineering fields, and this phenomenon is widely observed in nuclear power plants. Although turbulent structures play important roles in the velocity and pressure fields in a pipe, only a few studies have been conducted on the turbulent flow on an oscillating wall. In this study, direct numerical simulations were conducted to establish a large eddy simulation model for a turbulent flow on an oscillating wall and scrutinize the energy transfer between the grid scale (GS) and sub-grid scale (SGS). Although energy is generally transferred from the GS to SGS (forward scatter), it is likely that energy is transferred from the SGS to GS (backward scatter) under specific conditions. The present numerical results indicated that backward scatter exists in the production term in the case of a static wavy wall. On the other hand, such backward scatter could not be observed in the case of an oscillating wall. It is well known that separated flows and backward flows are generated behind the crest. Stronger backward flows accelerate the main flow and enhance the velocity gradients in a wide range behind the crest. In the case of an oscillating wall, the development of separated flow is immature because the shape of the wall is not fixed. Eventually, the backward scatter is deemed to be suppressed.

The Modular High Temperature Gas-cooled Reactor (MHTGR) could realize higher efficiency and lower costs by developing the multi-modular high temperature gas-cooled reactors combined with supercritical steam turbine unit. The coupling effects among different modules are crucial to the designs and operation analyses of the multi-modular reactors. By establishing the engineering simulator for multi-modular reactors, the coupling effects can be studied and optimized to advance the reactor designs, due to the advantages of real-time calculations and coupled calculations. As key energy transfer equipment, the steam generator is very important to the reactor operation, and focused in the modeling of the engineering simulator system for multi-modular reactors. In this paper, the once-through steam generator consisted of helical coils was modeled and optimized in the vPower integrated simulation platform. From the detailed analyses of the distributions of temperatures, heat flux, and other parameters along the heat transfer tubes, it showed that the steam generator model well presented the supercritical water properties and heat transfer characteristics inside helical tubes. Also, the heat transfer correlations of the supercritical water inside helical tubes were investigated, discussed and also compared to test the uncertainty and influence to the whole steam generator model. And the results indicated that most heat transfer correlations showed similar results and had little effect on the primary side in the steady state operation condition. In future work, the model and heat transfer characteristics of the supercritical steam generator will be further tested in more transients and integrated into complete engineering simulator for multi-module reactors.

Calcium silicate hydrate (CSH) is a main component of cement-based material required for constructing the geological repository. As in many countries, since the repository in Japan is constructed below water table, we must consider the interaction of radionuclide with cement materials altered around the repository after the backfill. Using fluorescence emission spectra, so far, the authors have investigated the interaction of Eu(III) (as a chemical analog of Am(III)) with CSH gels formed with no dried process, considering a condition saturated with groundwater. However, in such fluorescence emission behaviors, a deexcitation process of OH vibrators of light water and a quenching effect caused by Eu-Eu energy transfer between Eu atoms incorporated in the CSH gel must be considered. This study examined the fluorescence emission behavior of Eu(III) sorbed on CSH gels formed with no dried process, by using La(III) (non-fluorescent ions) as a diluent of Eu(III). Furthermore, the CSH samples were synthesized with CaO, SiO 2 , and heavy water (D 2 O) as a solvent in order to avoid the obvious deexcitation process of OH vibrators of light water. This study prepared CSH samples with the Ca/Si ratio set to 1.6, 1.0, and 0.5. A 1 mM solution of a given combination of Eu(III) and La(III) (Eu(III) content: 100%, 67%, 50% or 33%) was added into CSH gel sample. The contact time-period of the CSH gel with the Eu(III)/La(III) solution was set to 60 days. In the results, the peak around 618 nm was split into two peaks of 613 nm and 622 nm in the cases of Ca/Si = 1.0 and 1.6. Then, the peak of 613 nm decreased with increment of Eu(III)/La(III) ratio. This means that the relative intensity of 613 nm is useful to quantify the amount of Eu(III) incorporated in CSH gel. Besides, the intensity peak of 584 nm decayed with increment of Eu/La ratio, suggesting a quenching effect due to Eu-Eu energy transfer. However, the decay behavior of the fluorescence emission did not depend on the Eu/La concentration ratio. That is, such a quenching effect is neglectable. Additionally, the low Ca/Si ratio samples underwent slow attenuation of fluorescence and showed profiles similar to those of high Ca/Si ratio samples. Therefore, low Ca/Si ratio samples also include the reaction forming a complex on the surface of CSH gel with Eu(III). In other words, even if Ca/Si ratio is lower than 1.0, CSH gels would retard the migration of radionuclides released from the repository.

Fusion blanket is a key component for energy transformation and extraction in fusion reactor, in many kinds of blanket designs, the flow channel insert (FCI) made of a silicon carbide composite (SiCf/SiC) is a key element, which serves as electric and thermal insulator. In the high temperature liquid metal blanket (HTL) of fusion-based hydrogen production reactor (named FDS-III),. The multilayer flow channel inserts (MFCI) are put into the breeding zone to achieve the LiPb outlet temperature around 1000°C for high efficient production of hydrogen. However, the flow of liquid metal meandering through MFCI will cause strong magnetohydrodynamic (MHD) effect under the influence of the intense fusion magnetic field and the liquid metal in the neighboring channels will effect each other. When the FCI conductivity varies, the influence varies coresspondingly. So MHD effect of different FCI electrical conductivity is a key issue which should be concerned in HTL blanket. In this work, the numerical study aims at clarifying the MHD effect for liquid metal LiPb in the multi-layer flow channel inserts of different conductivity by using a code named MTC-H 2.0. The MHD flows with Special geometry of the HTL flow channels with MFCI were considered, velocity and FCI electrical conductivity’s effect on electromagnetic coupling are discussed.

In this study, the behavior of condensing single vapor bubble in subcooled boiling flow within two different vertical rectangular channels has been numerically investigated by using the VOF (Volume of Fluid) multiphase flow model. The mass and energy transfer model of bubble condensing process induced from the interfacial heat transfer was proposed to describe the interfacial transportation between the two phases. The results of VOF simulations showed good agreements with previous experimental data in the bubble size variation and lifetime. The bubble lifetime was almost proportional to bubble initial size and prolonged by system pressure. With the increase of subcooling, the bubble lifetime reduces significantly, and the effect of mass flux could be negligible. When bubble size increased, the bubble shape tends to be changed in a large-scale channel. The VOF simulation results of deformation have good agreement with those of Kamei’ experiment and the results of MPS (Moving Particle Semi-implicit) simulation in the large-scale channel. Furthermore, the initial bubble size, subcooling of liquid and system pressure play an important role to influence the bubble deformation behaviors significantly. The bubble could deform sharper with the increase of subcooling and initial diameter, or could breakup when the subcooling and the initial diameter reached a certain value at the last bubble stage. Whereas the trend of bubble deformation would be weaken with the increase of system pressure.

In this study, a thermal hydraulic behavior of the moderator in the CANDU reactor was numerically investigated by using CUPID code. KAERI has been developing a component-scale thermal hydraulics code, CUPID. The aim of the code is multi-dimensional, multi-physics and multi-scale thermal hydraulics analysis. This code adopts a three-dimensional, transient, two-phase and three-field model, and includes physical models and correlations of the interfacial mass, momentum, and energy transfer for the closure. To avoid the complexity to generate computational geometry around the matrix of 440 Calandria tubes, a porous media approach was applied. Flow resistance inside the porous media zone was derived from the empirical correlation of the frictional pressure loss. In order to consider the turbulent jet inflows from the inlet nozzles, the standard k-ε turbulence model was applied. For the grid dependency test, three different grid systems were tested. The moderator test vessel at Stern Laboratories Inc. (SLI) for the validation is a cylinder with a diameter of 2m and a length of 0.2m (a thin “slice” of CANDU-6 Calandria vessel). Since the axial flow is assumed to be invariant, two-dimensional calculation was performed. Vertical profile of the liquid temperature was compared with other calculation results as well as experimental data.

Two-fluid model can simulate two-phase flow by computational cost less than detailed two-phase simulation method such as interface tracking method. Therefore, two-fluid model is useful for thermal hydraulic analysis in large-scale domain such as rod bundles. However, since two-fluid model include a lot of constitutive equations, applicability of these constitutive equations must be verified by use of experimental results, and the two-fluid model has problems the result of analyses depends on accuracy of constitutive equations. To solve these problems, an advanced two-fluid model has been developed in Japan Atomic Energy Agency. In the model, an interface tracking method is combined with the two-fluid model to predict large interface structure behavior accurately. Liquid clusters and bubbles larger than a computational cell are calculated using the interface tracking method, and those smaller than a cell are simulated by the two-fluid model. Constitutive equations to evaluate the effect of small bubbles or droplets on two-phase flow required in the advanced two-fluid model as same as a conventional two-fluid model. However, dependency of small bubbles and droplets on two-phase flow characteristic is relatively small, and the experimental results to verify the equations are not required much. The turbulent dispersion force term is one of the most important constitutive equations for the advanced two-fluid model. The turbulent dispersion force term has been modeled by many researchers for the conventional two-fluid model. However, the existing models include effects of large bubbles and deformation of bubbles implicitly, these models are not applicable to the advanced two-fluid model. In this study, we develop the new model for turbulent dispersion force term. In this model, effect of large bubbles and deformation of bubbles are neglected. The liquid phase turbulent kinetic energy and bubble-induced turbulent kinetic energy are considered to evaluate driving force in the turbulent diffusion of small bubbles. The bubble-induced turbulent kinetic energy is given by the function of bubble diameter and local relative velocity, and the liquid phase turbulent kinetic energy is similar to the single phase flow case. Furthermore, we considered energy transfer from the bubble-induced kinetic energy to the liquid phase turbulent kinetic energy. To verify the developed model, the advanced two-fluid model and the model for turbulent dispersion term were incorporated to the 3-dimensional two-fluid model code ACE-3D, and comparisons between the results of analyses and air-water two-phase flow experiments in 200 mm diameter vertical pipe were performed.

The ANSI/ANS-6.4.3-1991 Standard, Gamma-Ray Attenuation Coefficients and Buildup Factors for Engineering Materials , is currently being updated by an American Nuclear Society (ANS) Working Group. The ANSI/ANS-6.4.3, 1991 standard, which is “withdrawn” due to the failure to meet the requirement of the ANS to have standards updated every ten years, contains buildup factor values that are derived from data that is over seventeen years old. In addition, computer technology has significantly improved since 1991, allowing for more complicated, computationally demanding codes to be utilized. Therefore, the ANSI/ANS-6.4.3-1991 standard is being re-visited to include updated data and to include modern codes. Gamma-ray buildup factors and attenuation coefficients are being generated for common shielding materials (e.g., concrete, steel, water, etc.) utilizing the ENDF/B-VI.8 cross-section data library, which is distributed by Brookhaven National Laboratory’s National Nuclear Data Center (NNDC). One modern code, Monte Carlo N-Particle (MCNP5/MCNPX), is being compared with PALLAS-1D (VII) and Anisotropic Source-Flux Iteration Technique (ASFIT-VARI), which were used to develop values included in ANSI/ANS-6.4.3-1991. PALLAS-1D (VII) is a code for direct integration of transport equation in one-dimensional plane and spherical geometries while ASFIT-VARI is a gamma-ray transport code system for one-dimensional finite systems. MCNP5 is a general purpose Monte Carlo radiation transport code that tracks particles (e.g., neutron and photons) at numerous energies in a three dimensional configuration of materials. The MCNP5 radiation transport code require response function input to provide dose and exposure output. Mass energy-absorption coefficients and mass energy-transfer coefficients are required to develop absorbed dose and exposure responses. The National Institute of Standards and Technology (NIST) provide and maintain the most up-to-date mass energy-absorption coefficient and mass energy-transfer coefficient database currently available. The NIST values are used in these initial buildup factor calculations to prove the validity of the methodology used and allow for preliminary comparisons. The energy absorption (dose in material) and exposure buildup factors are calculated at mfp values of interest, consistent with ANSI/ANS-6.4.3-1991 up to forty mfp. Comparisons between the new buildup factors and the previous results presented in the ANSI/ANS-6.4.3-1991 standard indicate that there is fairly good agreement. Differences in buildup factor values can be attributed to differences in cross-section data libraries, numerical methods, and physics treatments within the respective codes.

A component scale thermal hydraulic analysis code, CUPID (Component Unstructured Program for Interfacial Dynamics), is being developed for the analyses of components of a nuclear reactor, such as reactor vessel, steam generator, containment, etc. It adopts three-dimensional, transient, two-phase and three-field model, and includes various physical models and correlations of the interfacial mass, momentum and energy transfer for the closure relations of the two-fluid model. In the present paper, the two-phase models were assessed against the DOBO (DOwncomer BOiling) experiment, which was constructed to simulate the downcomer boiling phenomenon. It may happen in the downcomer of a nuclear reactor vessel during the reflood phase of a postulated loss of coolant accident. The stored energy release from the reactor vessel to the liquid inside the downcomer causes the boiling on the wall, and it can reduce the hydraulic head of the accumulated water, which is the driving force of water reflooding to the core. This phenomenon has been considered as a crucial safety issue of an advanced power reactor because it is concerned with the core cooling capability of the safety injection system. In this paper, the physical models and correlations that were incorporated into the CUPID code were introduced and the validation results against the experiment were reported. The benchmark calculation results concluded that the CUPID code can appropriately predict the boiling phenomena under a low pressure and low flow rate condition with modification of the bubble size correlation.

This paper presents recent advances in the development and validation of the two-phase flow topology models implemented in CFD-BWR, an advanced Computational Fluid Dynamics (CFD) computer code that allows the detailed analysis of the two-phase flow and heat transfer phenomena in Boiling Water Reactor (BWR) fuel assemblies under various operating conditions. The local inter-phase surface topology plays a central role in determining the mass, momentum, and energy exchanges between the liquid and vapor phases and between the two-phase coolant and the fuel pin cladding. The paper describes the topology map used to determine the local inter-phase surface topology and the role of the local topology in determining the inter-phase mass, momentum, and energy transfer. It discusses the relationship between the local interphase surface topology and the traditional channel flow regimes and presents results of experiment analyses in which computed local topologies are aggregated into flow regimes and compared with experimental observations.

Three-dimensional computer simulation and analyses of the horizontal steam generator thermal-hydraulics of the WWER 1000 nuclear power plant have been performed for 50% and 75% partial loads, 100% nominal load and 110% over-load. Presented results show water and steam mass flow rate vectors, steam void fraction spatial distribution, recirculation zones, swell level position, water mass inventory on the shell side, and other important thermal-hydraulic parameters. The simulations have been performed with the computer code 3D ANA, based on the “two-fluid” model approach. Steam-water interface transport processes, as well as tube bundle flow resistance, energy transfer, and steam generation within tube bundles are modelled with “closure laws”. Applied approach implies non-equilibrium thermal and flow conditions. The model is solved by the control volume procedure, which has been extended in order to take into account the 3D flow of liquid and gas phase. The methodology is validated by comparing numerical and experimental results of real steam generator operational conditions at various power levels of the WWER Novovoronezh, Unit 5. One-dimensional model of the horizontal steam generator has been built with the RELAP 5 standard code on the basis of the multidimensional two-phase flow structure obtained with the 3D ANA code. RELAP 5 and 3D ANA code results are compared, showing acceptable agreement.

This work reports on experimental studies to visualize nucleate boiling on the enhanced heat transfer surface of the hypervapotron for with application in the International Thermonuclear Experiment Reactor [ITER]. This research uses the simulant fluid Freon (R134A) instead of prototypic water to model the system performance. This results in much lower thermophysical conditions to represent the prototypic phenomena. By using reduced pressure, temperatures, etc, based on the critical physical properties of both working fluids, Freon and water, the dramatic drop in the level of these quantities with Freon allows the use of modest test conditions. The experiment was conducted for both saturated and subcooled boiling with different heat fluxes (from 50 to 300 kW/m 2 ). A comparison of the heat transfer performance of finned structures and flat surfaces were examined under particular fluid conditions. The uniqueness of this work is the visualization method that allows direct observation of the subcooled boiling process of the Hypervapotron surfaces. Working with a high speed (12,000 frames per second), high fidelity digital camera with variable magnifications (from 1×–25×), the sub-cooled boiling phenomena was observed in detail. A major conclusion of this work is the existence of two separate zones linked to different energy removal efficiency in hypervapotron. Under high heat flux condition, enhanced boiling heat transfer (about 20–30% higher than flat surface) was observed for hypervapotron effect, while saturated boiling happened in the cavity, and a large portion of the region was vapor filled. The process of vapor bubble rotation in the slot appeared to be helpful to enhance energy transfer, as evidenced by an improved wetting condition on the heating surfaces.

Westinghouse utilizes separate computer models for the loss-of-coolant accident (LOCA) mass and energy (M&E) release and containment response calculations for the containment design basis accident (DBA) analysis. The current LOCA M&E release model uses very conservative assumptions to calculate the rate of steam generator (SG) energy transfer to the reactor coolant system (RCS). Westinghouse is upgrading the containment DBA analysis to use W COBRA/TRAC ( W C/T) for the LOCA M&E release calculation and GOTHIC for the containment response calculation. Furthermore, the formerly separate calculations can now be performed concurrently to allow communication between the RCS and containment calculation models. W C/T utilizes more representative models for calculating the rate of SG energy transfer to the RCS than the currently approved LOCA M&E release calculation model. The full length emergency core heat transfer separate-effects and system-effects test (FLECHT-SEASET) data was used to validate the W C/T SG model for the post-blowdown LOCA M&E release calculation. The W C/T SG model originally calculated a higher SG energy transfer rate than the test data, which is conservative for containment analysis. The W C/T code was updated to include recommended changes to model the SG tube quench front. Following implementation of the recommended software changes, W C/T calculates a more realistic, yet still conservative, SG energy transfer rate. A W C/T plant model was developed to perform a LOCA M&E release calculation, and the W C/T results were compared to results from the currently approved LOCA M&E release calculation model. W C/T calculated a lower post-blowdown release rate than the currently approved LOCA M&E release calculation model. The lower post-blowdown release rate will provide additional peak pressure margin for plants that rely on passive cooling, containment spray, and/or fan coolers to control the containment peak pressure, and is also expected to extend the time of ice melt-out for ice condenser containments. The upgraded containment DBA analysis model that utilizes W C/T and GOTHIC concurrently will establish a conservative, but more representative long-term containment pressure and temperature response. Results based upon the GOTHIC containment model running with the W C/T plant model have shown a pressure reduction of approximately 10 psi (70 kPa) in the post-blowdown pressure transient.

The model subassembly of the BREST-type reactor core consists of a pin bundle of square arrangement. In this bundle there are two zones which differ in the pin diameter and heat production. The model pin bundle contains one spacer grid which is located near the mid-plane of the rod bundle geometry. The working is a eutectic alloy of 22% sodium (Na) plus 78% potassium (K). Three kinds of experiments were performed to observe the thermal and hydraulic behavior of the liquid metal coolant in the BREST core simulator. Results were obtained for the coolant exit temperature distribution, central measuring pin simulator external surface temperature distribution, and coolant velocity distribution over the perimeter of the measuring pin simulator. The experiments were performed five times with increasing pin power ratios. Analysis was performed on the model subassembly of the BREST-type reactor core using a subchannel analysis code MATRA and a computational fluid dynamics code CFX. Calculational results were compared against the experimental data. The experiment revealed that the temperature rise was strongly dependent upon the geometry of the pin simulator. In contrast to the experimental results, the MATRA results were mainly dependent upon the thermal and hydraulic conditions. It was concluded that MATRA requires modifications for the pressure drop correlations that were considered inappropriate for accurately simulating the coolant behavior near the BREST-type grid spacer. Hand calculations were additionally carried out under different assumptions to determine the coolant exit temperature distribution in the pin simulator. First, the hand calculation was performed to find the coolant exit temperature distribution assuming that there is no momentum or energy transfer between subchannels. Second, an assumption was made that the coolant mixing in the subchannel assembly took place instantaneously and the pressure was equilibrated at the channel exit. Since MATRA is based on a lumped parameter model, it only calculates the subchannel averaged velocity values. Here, CFX based on the finite volume method was utilized to calculate the velocity fields over the perimeter. Results from the experiment and CFX were averaged in each subchannel region so as to check on the tendency. The CFX analysis showed reasonable results which can be improved by imposing more detailed geometry accounting for the angle of the inclination of the grid spacer.

A new code, CFD-BWR [1], is being developed for the simulation of two-phase flow phenomena inside a BWR fuel bundle. These phenomena include coolant phase changes and multiple flow regimes which directly influence the coolant interaction with fuel assembly and, ultimately, the reactor performance. CFD-BWR is a specialized module built on the foundation of the commercial CFD code STAR-CD [2] which provides general two-phase flow modeling capabilities. New models describing the inter-phase mass, momentum, and energy transfer phenomena specific for BWRs have been developed and implemented in the CFD-BWR module. A set of experiments focused on two-phase flow and phase-change phenomena has been identified for the validation of the CFD-BWR code and results of two experiment analyses focused on the radial void distribution are presented. The close agreement between the computed results, the measured data and the correlation results provides confidence in the accuracy of the models.

During certain phases of a severe accident in a pressurized water reactor (PWR), the core becomes uncovered and steam carries heat to the steam generators through natural circulation. For PWR’s with U-tube steam generators and loop seals filled with water, a counter current flow pattern is established in the hot leg. This flow pattern has been experimentally observed and has been predicted using computational fluid dynamics (CFD). Predictions of severe accident behavior are routinely carried out using severe accident system analysis codes such as SCDAP/RELAP5 or MELCOR. These codes, however, were not developed for predicting the three-dimensional natural circulation flow patterns during this phase of a severe accident. CFD, along with a set of experiments at 1/7 th scale, have been historically used to establish the flow rates and mixing for the system analysis tools. One important aspect of these predictions is the counter current flow rate in the nearly 30 inch diameter hot leg between the reactor vessel and steam generator. This flow rate is strongly related to the amount of energy that can be transported away from the reactor core. This energy transfer plays a significant role in the prediction of core failures as well as potential failures in other reactor coolant system piping. CFD is used to determine the counter current flow rate during a severe accident. Specific sensitivities are completed for parameters such as surge line flow rates, hydrogen content, as well as vessel and steam generator temperatures. The predictions are carried out for the reactor vessel upper plenum, hot leg, a portion of the surge line, and a steam generator blocked off at the outlet plenum. All predictions utilize the FLEUNT V6 CFD code. The volumetric flow in the hot leg is assumed to be proportional to the square root of the product of normalized density difference, gravity, and hydraulic diameter to the 5 th power. CFD is used to determine the proportionality constant in the range from 0.11 to 0.13 and termed a discharge coefficient. The value is relatively unchanged for typical surge line flow rates as well as the hydrogen content in the flow. Over a significant range of expected temperature differences for the steam generator and reactor vessel upper plenum, the discharge coefficient also remained consistent. The discharge coefficient is a suitable model for determining the hot leg counter current flow rates during this type of severe accident.