The convective heat transfer of rod bundle flow with spacer grid was investigated preliminarily for nuclear reactor core application. As the test fluid, the water was used. To simulate the nuclear fuel assembly, 4×4 rod bundle with P/D (=pitch between rods/rod diameter) of ∼1.35 was prepared together with a spacer grid with twist-mixing vane. A single heated section with five thermocouples embedded in the surface along the circumferential direction was installed around the center subchannel. The measurements of wall temperatures were carried out upstream and downstream of spacer grid. For the rod bundle flow at the inlet of spacer grid (i.e., upstream of spacer grid), the wall temperatures at the gap and subchannel centers exhibited the higher and lower, respectively, which was because in the subchannel center, the axial flow velocity became higher, as compared with the gap center. On the other hand, downstream of spacer grid, the rod wall toward the tip of twist-mixing vane showed the lowest temperature in the measurements along the circumferential direction of rod wall. Near the twist-mixing vane, the averaged wall temperature was observed to be remarkably low, which implies that the twist-mixing vane is an effective tool to enhance the convective heat transfer performance. However, along the axial flow direction behind the spacer grid, the averaged wall temperatures became to increase, and the enhancement of convective heat transfer performance by mixing vane faded away.

The transient pool boiling heat transfer of zircaloy and stainless steel rodlets was investigated using a quenching method. The influences of quenchant temperature, test specimen size, and material property were examined. The groove-structured surface was also prepared and tested, which could be a valuable attempt to study the effect of a machined surface structure of a vertical rodlet on the rapid cooling heat transfer performance. A test specimen with a small volume-to-heat transfer surface area and small heat capacity showed a shorter quenching time. In the low quenchant temperature condition, a stable film boiling regime was likely to be hardly observed. The grooved surface appeared to have a shorter quenching duration, which could be due to its small volume-to-area ratio. In addition, the groove-structured portion seemed to help the liquid make contact with the heated wall earlier than the plain portion. Based on this work, it was found that the quenching duration and the position of vapor film rupture can be controlled by the surface structure.

Experimental and CFD (Computational Fluid Dynamics) analyses were conducted to determine the pressure loss coefficient of a spacer grid for a tight-lattice rod bundle simulating a dual-cooled annular fuel (DCAF) assembly. The DCAF is designed to have both internally and externally cooled channels by adopting annular pellet and dual claddings. The Korea Atomic Energy Research Institute proposed the DCAF assembly for the Korean optimum power reactor (OPR1000) which is a 12×12 square rod bundle with a rod pitch-to-diameter ratio (P/D) of 1.08. The pressure loss across the spacer grid specially designed for the 12×12 rod bundle is required in order to determine the inner to outer cooling channel flow split. Hence, the pressure drop of the spacer grid was measured using the full-size rod bundle for a Reynolds number ranging from 2e+04 to 2e+05. A CFD analysis was also performed to predict the pressure drop of the spacer grid used in the full-size bundle experiment. Only a single grid span of the test rod bundle was modeled in this CFD simulation. The grid loss coefficients by both the experiment and CFD analysis was shown to decrease as the Reynolds number increases. The measured loss coefficient was estimated from 1.30 to 1.50 for the normal operating condition of the reactor core with the DCAF, i.e., Re=3.5e+05. The CFD simulation predicted the grid loss coefficient being approximately 5% higher than the measured value.

In order to ensure the compactness and high-power density of a nuclear power reactor, the research on tight-lattice fuel bundle, with a narrow gap distance between fuels, has been highlighted. Recently, KAERI (Korea Atomic Energy Research Institute) has been developing dual-cooled annular fuel to increase a significant amount of the reactor power in OPR1000 (Optimized Power Reactor), a PWR (Pressurized Water Reactor) optimized in the Republic of Korea. The dual-cooled annular fuel is configured to allow a coolant flow through the inner channel as well as the outer channel. To introduce the dual-cooled annular fuel to OPR1000 is aiming at increasing the reactor power by 20% and reducing the fuel-pellet temperature by 30%, as compared to the cylindrical solid fuel, without a change in reactor components. In such a case, due to larger outer diameter of a dual-cooled annular fuel, the dual-cooled annular fuel assembly exhibits a smaller P/D (Pitch-to-Diameter ratio) than the conventional cylindrical solid fuel assembly. In other words, the dual-cooled annular fuel array becomes the tight-lattice fuel bundle configuration, and such a change in P/D can significantly affect the thermal-hydraulic characteristics in nuclear reactor core. In this paper, the pressure drop and flow pulsation in tight-lattice rod bundle were investigated. As the test sections, the tight-lattice rod bundle of P/D = 1.08 was prepared with the regular rod bundle of P/D = 1.35. The friction factors in P/D = 1.08 appeared smaller than those in P/D = 1.35. For P/D = 1.08, the twist-vane spacer grid became the larger pressure loss coefficients than the plain spacer grid. In P/D = 1.08, the flow pulsation, quasi-periodic oscillating flow motion, was visualized successfully by PIV (Particle Image Velocimetry) and MIR (Matching Index of Refraction) techniques. The peak frequency and power spectral density of flow pulsation increased with increasing the Reynolds number. Our belief is that this work can contribute to a design of nuclear reactor with tight-lattice fuel bundle for compactness and power-uprate and a further understanding of the coolant mixing phenomena in a nuclear fuel assembly.

Turbulent air flow through subchannels is simulated to obtain characteristic frequency on large scale flow pulsation in square arrayed 6-rod bundles. The turbulent mixing between the subchannels is dominated by the large scale flow pulsation phenomenon. Therefore, it is crucial to understand flow patterns in the subchannels for the design of reliable nuclear reactor elements. Numerical simulation using a CFD code is performed by solving the unsteady equations with large eddy simulation (LES) model. For the numerical simulation, two kinds of grid including two subchannels are created. Models with two different P/D ratios (P/D = 1.08, 1.35) are simulated. P/D = 1.08 case is our main concern to investigate flows in tight lattice rod buddle. The shortened channel length with periodic boundary condition is employed to reduce the computing time. The geometry of the simulated model is based on the 3-D experimental equipment. The velocity fluctuation data in the stream-wise and cross directions are obtained from the simulated model to find the characteristic frequency. The large scale flow pulsations are explained by time history of velocity, Fast Fourier Transform analysis and cross-correlation analysis. The fluctuation frequency at the center of gap and 20 mm away from that point are obtained. The characteristic frequency is 87.5 Hz through the entire monitoring points. Results demonstrate that the periodic flow pulsations occur with small pitch-to-diameter ratio (P/D = 1.08) and the vortices trains are highly synchronized. There is no large scale pulsation in P/D = 1.35 case.

A CFD study was performed to simulate the steady-state void distribution benchmark based on the NUPEC PWR Subchannel and Bundle Tests (PSBT). The void distribution benchmark provides measured void fraction data over a wide range of geometrical and operating conditions in a single subchannel and fuel bundle. This CFD study simulated the boiling flow in a single subchannel. A CFD code was used to predict the void distribution inside the single subchannel. The multiphase flow model used in this CFD analysis was a two-fluid model in which liquid (water) and vapor (steam) were considered as continuous and dispersed fluids, respectively. A wall boiling model was also employed to simulate bubble generation on a heated wall surface. The CFD prediction with a small diameter of vapor bubble shows a higher void fraction near the heated wall and a migration of void in the subchannel gap region. A measured CT image of void distribution indicated a locally higher void fraction near the heated wall for the test conditions of a subchannel averaged void fraction of less than about 20%. The CFD simulation predicted a subchannel averaged void fraction and fluid density which agree well with the measured ones for a low void condition.

A simplified thermal-hydraulic program was developed to quickly calculate the minimum Departure from a Nucleate Boiling Ratio (DNBR) in a pressurized water reactor (PWR) core which is a measure for the core thermal margin. The accuracy of the simplified thermal-hydraulic program would largely depend on the size of the axial nodal sections particularly for the non-iterative scheme for solving the transport coefficient conservation equations. The number of axial nodes in the faster-running DNBR program varies depending on its applications to a PWR core monitoring and protection systems and a reactor safety analysis. It is therefore important to examine the effects of the axial nodal sections in the simplified DNBR program for its application to an advanced PWR core protection system. This paper presents the uncertainties of the minimum DNBR (MDNBR) and the DNBR margin assessment depending on the number of axial nodes. The DNBR margin significantly decreases due to very large MDNBR uncertainty for the number of axial nodes of less than 20.

This CFD study is to simulate a coolant (gas) flow and heat transfer in a PBR core during a normal operation. This study used a pebble array with direct area contacts among the pebbles which is one of the pebbles arrangements for a detailed simulation of PBR core CFD studies. A CFD model is developed to more adequately represent the pebbles randomly stacked in the PBR core. The CFD predictions showed a large variation of the temperature on the pebble surface as well as in the pebble core. The temperature drop in the outer graphite layer is smaller than that in the pebble-core region. This is because the thermal conductivity of graphite is higher than the fuel (UO 2 mixture) conductivity in the pebble core. Higher pebble surface temperature is predicted downstream of the pebble contact due to a reverse flow. Multiple vortices are predicted to occur downstream of the spherical pebbles due to a flow separation. The coolant flow structure and fuel temperature in the PBR core appears to largely depend on the in-core distribution of the pebbles.

An experiment was performed in a wind tunnel to investigate the flow structure in a rod bundle with a hybrid vane grid. The hybrid vane is a flow-mixing device, which consists of two pairs of primary and secondary vanes in a cell. The test section is a rectangular channel (300 mm × 300 mm × 2400 mm) including 3 × 3 rod (75 mm diameter) array with a spacer grid. The pitch to diameter ratio of the rod array is 1.33. The flow structures downstream the grid are measured at Reynolds number of 1.2 × 10 5 for 35-degree deflecting angle of the hybrid flow-mixing vane. The data are obtained for the distributions of the time mean axial velocity, lateral velocity, and turbulent intensities in 3 component directions over a center subchannel along axial locations and compared with the previous results of split vane grid that has two vanes in a cell. The results show that the mixing efficiency of the hybrid vane grid could be similar with that of the split vane grid because swirl factor of the hybrid vane grid is higher than that of split vane grid and the magnitude of axial turbulent intensity, turbulent diffusion coefficient, and cross flow factor is similar to each other in spite of differences of the vane numbers and shape in a cell between hybrid and split vane grids.

The spacer grid is one of the structural components for the fuel assembly. In order to increase or extend the fuel life cycle, a spacer grid which has a much higher performance from the thermal/hydraulic and mechanical/structural point of view will be needed. From the thermal/hydraulic viewpoint, the CHF margin is very important in order to extend its life. Particularly, the mixing flow or cross flow between the subchannels have to be reinforced for this purpose. From the mechanical/structural viewpoint, the critical strength and the fuel rod support behaviour of a spacer grid are the same as the TH performance improvement for the next generation fuel. A computational fluid dynamics (CFD) analysis was performed to investigate the coolant mixing in a nuclear fuel bundle that is promoted by the mixing vane on the grid spacer. Single and multiple subchannels of one grid span of the fuel bundle were modeled to simulate a 5by5 rod array experiment with the mixing vane. The three-dimensional CFD models were generated by a structured multi-block method. The standard k-ε turbulence model was used in the current CFD simulation since it is practically useful and converges well for the complex turbulent flow in a nuclear fuel bundle. The CFD predictions of the axial and lateral mean flow velocities showed a somewhat larger difference from the experimental results near the spacer but represented the overall characteristics of the coolant mixing well in a nuclear fuel bundle with the mixing vane. Comparison of the single and multiple subchannel predictions shows a good agreement for the flow characteristics in the central subchannel of the rod array. The simulation of the multiple subchannels shows a slightly off-centered swirl in the peripheral subchannels due to the external wall of the rod array. It also shows no significant swirl and crossflow in the wall subchannels and the corner subchannels. In addition to this, the impact and the stress analysis of a spacer grid are accomplished by the FE method. The FE model was created using I-DEAS [4], and the ABAQUS/explicit version 6.3 commercial code was used for the solver. The FE analysis procedure was established, the FE analyses results were verified by the experimental method. The developed spacer grid will be evaluated from the thermal/hydraulic and mechanical/structural design criteria.

The numerical predictions using the standard and RNG k–ε eddy viscosity models, differential stress model (DSM) and algebraic stress model (ASM) are examined for the turbulent flow in a nuclear fuel bundle with the mixing vane. The hybrid (first-order) and curvature-compensated convective transport (CCCT) schemes were used to examine the effect of the differencing scheme for the convection term. The CCCT scheme was found to more accurately predict the characteristics of turbulent flow in the fuel bundle. There is a negligible difference in the prediction performance between the standard and RNG k-ε models. The calculation using ASM failed in meeting the convergence criteria. DSM appeared to more accurately predict the mean flow velocities as well as the turbulence parameters.

A computational fluid dynamics (CFD) analysis was performed to investigate the coolant mixing in a nuclear fuel bundle that is promoted by the mixing vane on the grid spacer. Single and multiple subchannels of one grid span of the fuel bundle were modeled to simulate a 5×5 rod array experiment with the mixing vane. The three-dimensional CFD models were generated by a structured multi-block method. The standard k-ε turbulence model was used in the current CFD simulation since it is practically useful and converges well for the complex turbulent flow in a nuclear fuel bundle. The CFD predictions of axial and lateral mean flow velocities showed a somewhat large difference from the experimental results near the spacer but represented the overall characteristics of coolant mixing well in a nuclear fuel bundle with the mixing vane. Comparison of single and multiple subchannel predictions shows good agreement of the flow characteristics in the central subchannel of the rod array. The simulation of multiple subchannels shows a slightly off-centered swirl in the peripheral subchannels due to the external wall of the rod array. It also shows no significant swirl and crossflow in the wall subchannels and the corner subchannels.