Search

The ITER Organization (IO) developed a thermal-hydraulic (TH) model of the complex first wall and blanket (FW/BLK) cooling channels to determine gas flow rate and pressure required to effectively blow out the water in the FW/BLK. In addition, US ITER conducted experiments for selected geometries of FW/BLK flow channels to predict the blowout parameters. The analysis indicates that as low as 2 MPa of pressure difference over the blanket modules will ensure substantial evacuation of the water in blankets with just a few percent remaining in the blanket flow channels. A limited validation study indicates that the analysis yields less conservative results to compare against data collected from experiments. Therefore, the designed blow out flow of the drying system was selected with a large margin above the measured values to ensure the blow out operation.

The nuclear industry is increasingly considering cartridge-style experiments for irradiation testing of advanced reactor fuels and materials under flowing conditions. Cartridge loops do not require the extensive support infrastructure that are necessary for external flow loops and minimize the possibility of coolant solidification over the long distance from the reactor to the external facilities. However, there is a general lack of quality flow data for internally heated fluids in an annular configuration representative of a cartridge-type irradiation experiment, particularly one with natural circulation. To address this data need, a series of experiments was conducted to measure the natural circulation flow rates of pressurized water in a sealed, internally heated vessel with annular flow conditions that represent a molten salt or sodium cartridge loop. Temperatures and flow rates were measured under steady-state and transient conditions. This paper describes the facility, methods, and results of the experiments, including the determination of nondimensional parameters. A simple 1D model of the natural convection flow rates agrees well with the experimental results. Applying this model to simulate a liquid salt cartridge experiment predicts that natural circulation flow might be able to provide liquid salt Reynolds numbers similar to those of some molten salt reactor concepts at relevant power densities.

Populations of small helium gas bubbles were introduced into a flowing mercury experiment test loop to evaluate mitigation of beam-pulse induced cavitation damage and pressure waves. The test loop was developed and thoroughly tested at the Spallation Neutron Source (SNS) prior to irradiations at the Los Alamos Neutron Science Center - Weapons Neutron Research Center (LANSCE-WNR) facility. Twelve candidate bubblers were evaluated over a range of mercury flow and gas injection rates by use of a novel optical measurement technique that accurately assessed the generated bubble size distributions. Final selection for irradiation testing included two variations of a swirl bubbler provided by Japan Proton Accelerator Research Complex (J-PARC) collaborators and one orifice bubbler developed at SNS. Bubble populations of interest consisted of sizes up to 150 m in radius with achieved gas void fractions in the 10^-5 to 10^-4 range. The nominal WNR beam pulse used for the experiment created energy deposition in the mercury comparable to SNS pulses operating at 2.5 MW. Nineteen test conditions were completed each with 100 pulses, including variations on mercury flow, gas injection and protons per pulse. The principal measure of cavitation damage mitigation was surface damage assessment on test specimens that were manuallymore » replaced for each test condition. Damage assessment was done after radiation decay and decontamination by optical and laser profiling microscopy with damaged area fraction and maximum pit depth being the more valued results. Damage was reduced by flow alone; the best mitigation from bubble injection was between half and a quarter that of flow alone. Other data collected included surface motion tracking by three laser Doppler vibrometers (LDV), loop wall dynamic strain, beam diagnostics for charge and beam profile assessment, embedded hydrophones and pressure sensors, and sound measurement by a suite of conventional and contact microphones.« less

Because polyaromatic hydrocarbons show high thermal stability, an example of these compounds, phenylnaphthalene, was tested for solar thermal-power applications. Although static thermal tests showed promising results for 1-phenylnaphthalene, loop testing at temperatures to 450 ℃ indicated that the fluid isomerized and degraded at a slow rate. In a loop with a temperature high enough to drive the isomerization, the higher melting point byproducts tended to condense onto cooler surfaces. This would indicate that the internal channels of cooler components of trough solar electric generating systems, such as the waste heat rejection exchanger, may become coated or clogged affecting loop performance. Thus, pure 1-phenylnaphthalene, without addition of stabilizers, does not appear to be a fluid that would have a sufficiently long lifetime (years to decades) to be used in a loop at temperatures significantly greater than the current 400 ℃ maximum for organic fluids. Similar degradation pathways may occur with other organic materials. The performance of a concentrating solar loop using high temperature fluids was modeled based on the National Renewable Laboratory Solar Advisory Model. It was determined that a solar-to-electricity efficiency of up to 30% and a capacity factor of 60% could be achieved using a high efficiency collector and 12 h thermal energy storage when run at a field outlet temperature of 550 ℃.

The Spallation Neutron Source (SNS) at the Oak Ridge National Laboratory has been conducting RD and protective gas walls that isolate the vessel from the damaging effects of collapsing cavitation bubbles. The latter has demonstrated good damage mitigation during in-beam testing with limited pulses, and adequate gas wall coverage at the beam entrance window has been demonstrated with the SNS mercury target flow configuration using a full scale mercury test loop. A question on the required area coverage remains which depends on results from SNS target post irradiation examination. The small gas bubble technique has been less effective during past in-beam tests but those results were with un-optimized and un-verified bubble populations. Another round of in-beam tests with small gas bubbles is planned for 2011. The first SNS target was removed from service in mid 2009 and samples were cut from two locations at the target’s beam entrance window. Through-wall damage was observed at the innermost mercury vessel wall (not a containment wall). The damage pattern suggested correlation with the local mercury flow condition which is nearly stagnant at the peak damage location. Detailed post irradiation examination of the samples is under way that will assess the erosion and measure irradiation-induced changes in mechanical properties. Similar samples were cut from the second SNS target after it was removed from service in mid 2010. More extensive damage was observed on the target inner wall but damage to the containment wall was minimal.

Proton beam experiments investigating cavitation damage in short pulse mercury spallation targets were performed at LANSCE – WNR in July of 2008. They included two main areas for investigation: damage dependence on mercury velocity using geometry more prototypic to the SNS target than previously employed and damage dependence on incident proton beam flux intensity. The flow dependence experiment employed six test targets with mercury velocity in the channel ranging from 0 to more than 4 m/s. Each was hit with 100 WNR beam pulses with peak proton flux equivalent to that of SNS operating at 2.7 MW. Damage dependence on incident proton beam flux intensity was also investigated with three intensity levels used on simple rectangular shaped targets without mercury flow. Intensity variation was imposed by focusing the beam differently while maintaining protons per pulse. This kept total energy deposited in each target constant. A fourth test target was hit with various beams: constant protons and varied spot size; constant spot size and varied protons. No damage will be assessed in this case. Instead, acoustic emissions associated with cavitation collapse were measured by laser Doppler vibrometer (LDV) from readings of exterior vessel motions as well as by mercury wetted acoustic transducers. This paper will provide a description of the experiment and present available results. Damage assessment will require several months before surface analysis can be completed and was not available in time for IWSMT-9.

Establishment of a gas layer between the flowing liquid and container wall is proposed for mitigating the effects of cavitation in mercury spallation targets. Previous work has shown an order of magnitude decrease in damage for a gas layer developed in a stagnant mercury target for an in-beam experiment. This work is aimed at extending these results to the more complex conditions introduced by a flowing mercury target system. A water-loop has been fabricated to provide initial insights on potential gas injection methods into a flowing liquid. An existing full-scale flow loop designed to simulate the Spallation Neutron Source target system will be used to extend these studies to mercury. A parallel analytical effort is being conducted using computational fluid dynamics (CFD) modeling to provide direction to the experimental effort. Some preliminary simulations of gas injection through a single hole have been completed and show behavior of the models that is qualitatively meaningful.

Computational Fluid Dynamic (CFD) numerical simulations were performed for the flow inside the Spallation Neutron Source jet-flow target vessel at Oak Ridge National Laboratory. Different flow rates and beam conditions were tested to cover all the functioning range of the target, but for brevity, only the nominal case with a mass flow rate of 185 kg/s and a beam power of 1.54MW is presented here. The heat deposition rate from the proton beam was computed using the general-purpose Monte Carlo radiation transport code MCNPX and the commercial CFD code ANSYS-CFX was used to determine the flow velocity in the mercury and the temperature fields in both the mercury and the stainless steel vessel. Boundary conditions, turbulence model and mesh effects are presented in depth. To validate the numerical approach, Particle Imagery Velocimetry (PIV) measurements on a water-loop setup with an acrylic jet-flow target mock-up were performed and compared to the numerical simulations. It was found that a sustained wall jet was developed across the whole length of the vulnerable surface, confirming the good design of the jet-flow target. Overall, good agreements were observed between the experiments and the simulations: the velocity contours and the shape of the recirculation zone near the side baffle are qualitatively similar. However, some differences were also observed that underlines the shortcomings of both the numerical simulations and the experimental measurements.Copyright © 2015 by ASME

Intense proton beam-induced heating of the spallation neutron source mercury target will cause pressure spikes that lead to the formation of cavitation bubbles in the mercury. Erosion of the mercury container walls caused by violent collapse of bubbles could potentially limit its service lifetime. In-beam tests for a limited number of pulses (

Previous experiments to examine corrosion resulting from thermal gradient mass transfer of type 316L stainless steel in mercury were conducted in thermal convection loops (TCLs) with an Hg velocity of about 1 m/min. These tests have now been supplemented with a series of experiments designed to examine the influence of increased flow velocity and possible cavitation conditions on compatibility. In one experiment, the standard TCL design was modified to include a reduced section in the hot leg that provided a concomitant increase in the local velocity by a factor of five. In addition, a pumped-loop experiment was operated with a flow velocity of about 1 m/s. Finally, a TCL was modified to include an ultrasonic transducer at the top of the hot leg in an attempt to generate cavitation conditions with corresponding extreme local velocity associated with collapsing bubbles. The results indicate that compatibility of type 316L/316LN stainless steel does not depend significantly on liquid metal velocity in the range of 1 m/min to 1 m/s. Benchtop cavitation experiments revealed susceptibility of 316L coupons to significant weight losses and increases in surface roughness as a result of 24 h exposure to 1.5 MPa pressure waves in Hg generated ultrasonically at 20 kHz. However, attempts to generate cavitation conditions on coupons inside the TCL with the ultrasonic transducer proved largely unsuccessful.

The heat transfer coefficient is inferred from measurements for mercury flowing in a channel of cross-section 2 mm×40 mm with flow velocities from 1 m/s to 4 m/s and heat fluxes from 192 kW/m2 to 1.14 MW/m2. Mercury bulk temperatures vary from 67°C to 143°C. Inferred heat transfer coefficients agree with open literature tube data when compared on a Nusselt versus. Peclet number plot, with Nusselt numbers examined from 8 to 17 and Peclet numbers examined from 790 to 3070.

Pressure waves created in liquid mercury targets at the pulsed Spallation Neutron Source (SNS) at Oak Ridge National Laboratory induce cavitation damage on the stainless steel target vessel. The cavitation damage is thought to limit the lifetime of the target for power levels at and above 1 MW. Severe through-wall cavitation damage on an internal wall near the beam entrance window has been observed in spent-targets. Surprisingly though, there is very little damage on the walls that bound an annular mercury channel that wraps around the front and outside of the target. The mercury flow through this channel is characterized by smooth, attached streamlines. One theory to explain this lack of damage is that the uni-directional flow biases the direction of the collapsing cavitation bubble, reducing the impact pressure and subsequent damage. The theory has been reinforced by in-beam separate effects data. For this reason, a second-generation SNS mercury target has been designed with an internal wall jet configuration intended to protect the concave wall where damage has been observed. The wall jet mimics the annular flow channel streamlines, but since the jet is bounded on only one side, the momentum is gradually diffused by the bulk flow interactions as it progresses around the cicular path of the target nose. Numerical simulations of the flow through this jet-flow target have been completed, and a water loop has been assembled with a transparent test target in order to visualize and measure the flow field. This paper presents the wall jet simulation results, as well as early experimental data from the test loop.Copyright © 2014 by ASME

The Advanced Neutron Source (ANS) reactor was designed to provide a research tool with capabilities beyond those of any existing reactors. One portion of its state-of-the-art design required high velocity fluid flow through narrow channels between the fuel plates in the core. Experience with previous reactors had shown that fuel plate damage could occur if debris became lodged at the entrance to these channels. Such debris disrupts the fluid flow to the plate surfaces and could prevent adequate cooling of the fuel. Preliminary ANS designs addressed this issue by providing an unheated entrance length for each fuel plate so that any flow disruption would have time to recover before reaching the heated portions of the fuel plates further downstream. As part of the safety analysis, the adequacy of this unheated entrance length was assessed using both analytical models and experimental measurements. The Flow Blockage Test Facility (FBTF) was designed and built to conduct experiments in an environment closely matching the ANS channel geometry. The FBTF permitted careful measurements of both heat transfer and hydraulic parameters. In addition to these experimental efforts, a thin rectangular channel was modeled using the FLUENT computational fluid dynamics computer code. The numerical results were compared to the experimental data to benchmark the hydrodynamics of the model. After this comparison, the model was extended to include those elements of the safety analysis difficult to measure experimentally. These elements included the high wall heat flux pattern and variable fluid properties.

Pressure waves created in liquid mercury pulsed spallation targets have been shown to create cavitation damage to the target container. One way to mitigate such damage would be to absorb the pressure pulse energy into a dispersed population of small bubbles, however, creating such a population in mercury is difficult due to the high surface tension and particularly the non-wetting behavior of mercury on gas-injection hardware. If the larger injected gas bubbles can be broken down into small bubbles after they are introduced to the flow, then the material interface problem is avoided. Research at the Oak Ridge National Labarotory is underway to develop a technique that has shown potential to provide an adequate population of small-enough bubbles to a flowing spallation target. This technique involves gas injection at an orifice of a geometry that is optimized to the turbulence intensity and pressure distribution of the flow, while avoiding coalescence of gas at injection sites. The most successful geometry thus far can be described as a square-toothed orifice having a 2.5 bar pressure drop in the mercury flow of 8 L/s for one of the target inlet legs. High-speed video and high-resolution photography have been used to quantify the bubble population on the surface of the mercury downstream of the gas injection site. Also, computational fluid dynamics has been used to optimize the dimensions of the toothed orifice based on a RANS computed mean flow including turbulent energies such that the turbulent dissipation and pressure field are best suited for turbulent break-up of the gas bubbles.Copyright © 2010 by ASME

The Spallation Neutron Source (SNS) is an accelerator-based neutron source at Oak Ridge National Laboratory (ORNL).The nuclear spallation reaction occurs when a proton beam hits liquid mercury. This interaction causes thermal expansion of the liquid mercury which produces high pressure waves. When these pressure waves hit the target vessel wall, cavitation can occur and erode the wall. Research and development efforts at SNS include creation of a vertical protective gas layer between the flowing liquid mercury and target vessel wall to mitigate the cavitation damage erosion and extend the life time of the target. Since mercury is opaque, computational fluid dynamics (CFD) may be used as a diagnostic tool to visualize the behavior of the liquid mercury and guide the experimental efforts. In this study, CFD simulations of three dimensional, unsteady, turbulent, two-phase flow of helium gas injection in flowing liquid mercury over smooth, vertically grooved and horizontally grooved walls are carried out with the commercially available CFD code Fluent-12 from ANSYS. The Volume of Fluid (VOF) model is used to track the helium-mercury interface. V-shaped vertical and horizontal grooves with 0.5 mm pitch and about 0.7 mm depth were machined in the transparent wall of acrylic test sections. Flow visualization data of helium gas coverage through transparent test sections is obtained with a high-speed camera at the ORNL Target Test Facility (TTF). The helium gas mass flow rate is 8 mg/min and introduced through a 0.5 mm diameter port. The inlet mercury mass flow rate is 51 kg/s and the predicted local mercury velocity is 0.9 m/s. In this paper, the helium gas flow rate and the local mercury velocity are kept constant for the three cases. Time integration of predicted helium gas volume fraction over time is done to evaluate the gas coverage and calculate the average thickness of the helium gas layer. The predicted time-integrated gas coverage over vertically grooved and horizontally grooved test sections is better than over a smooth wall. The simulations show that the helium gas is trapped inside the grooves.Copyright © 2009 by ASME

The Spallation Neutron Source (SNS) facility in Oak Ridge, Tennessee uses a liquid mercury target that is bombarded with protons to produce a pulsed neutron beam for materials research and development. In order to mitigate expected cavitation damage erosion (CDE) of the containment vessel, a two-phase flow arrangement of the target has been proposed and was earlier proven to be effective in significantly reducing CDE in non-prototypical target bodies. This arrangement involves covering vulnerable surfaces with a protective layer of gas. The difficulty lies in establishing a persistent gas layer that is oriented vertically and holds up to the strong buoyancy force and the turbulent mercury flow. Several new multiphase experiments have been completed at the Oak Ridge National Laboratory toward developing such layers. The gas hold-up is accomplished by machining regular features (grooves or pits) into the wall with dimensions on the order of 1 mm. The thickness of the gas layer varies, and it is currently unknown how thick a layer must be in order to successfully mitigate the damage, although this aspect is also under investigation. The paper includes a description of the various tests, a presentation of high-speed video images of the gas/mercury interaction viewed through a transparent window, and a discussion of how the results can be used to design a new SNS target that might be resistant to cavitation damage erosion.Copyright © 2009 by ASME

The corrosion of aluminum alloy fuel cladding in ultrahigh flux research reactors such as the Advanced Neutron Source may pose special temperature and structural problems associated with the buildup of reaction products during the lifetime of the core. In order to ensure that both the fuel and cladding integrity are maintained for the special conditions to be imposed on the ANS fuel plates, an experimental facility was constructed that permits the examination of material behavior for a wide range of thermal-hydraulic conditions, particularly high heat flux (to 20 MW/m{sup 2}) and coolant velocity (to 35m/sec). Tests have been conducted on specimens of 6061 aluminum that yielded meaningful comparisons to earlier data (i.e., High Flux Isotope Reactor and Advanced Test Reactor) and illustrated several important aspects of the process. In addition to the direct measurements of temperature response, these tests indicated that the film growth rate and morphology were sensitive to the heat flux and coolant water chemistry, as well as other system temperatures. The corrosion products were mainly boehmite (Al{sub 2}O{sub 3}{center dot}H{sub 2}O) and an iron-rich barrier-type layer on the outer surface of the boehmite, which appeared to form in certain instances as a consequence of mass transport frommore » the cooler sections of the loop. Eventually, also depending upon the particular coolant and heat-transfer conditions, the films underwent spallation; and this event was followed by an aggressive internal reaction in the alloy underlying the film.« less

The Spallation Neutron Source (SNS) facility in Oak Ridge, Tennessee uses a liquid mercury target that is bombarded with protons to produce a pulsed neutron beam for materials research and development. In order to mitigate expected cavitation damage erosion (CDE) of the containment vessel, a two-phase flow arrangement of the target has been proposed and was earlier proven to be effective in significantly reducing CDE in non-prototypical target bodies. This arrangement involves covering the beam “window”, through which the high-energy proton beam passes, with a protective layer of gas. The difficulty lies in establishing a stable gas/liquid interface that is oriented vertically with the window and holds up to the strong buoyancy force and the turbulent mercury flow field. Three approaches to establishing the gas wall have been investigated in isothermal mercury/gas testing on a prototypical geometry and flow: (1) free gas layer approach, (2) porous wall approach, and (3) surface-modified approach. The latter two of these approaches show success in that a stabilized gas layer is produced. Both of these successful approaches capitalize on the high surface energy of liquid mercury by increasing the surface area of the solid wall, thus increasing gas hold up at the wall. In this paper, a summary of these experiments and findings is presented as well as a description of the path forward toward incorporating the stabilized gas layer approach into a feasible gas/mercury SNS target design.Copyright © 2008 by ASME

Investigations in the area of two-phase flow at the Oak Ridge National Laboratory’s (ORNL) Spallation Neutron Source (SNS) facility are progressing. It is expected that the target vessel lifetime could be extended by introducing gas into the liquid mercury target. As part of an effort to validate the two-phase computational fluid dynamics (CFD) model, simulations and experiments of gas injection in stagnant water have been completed. The volume of fluid (VOF) method as implemented in ANSYS-CFX was used to simulate the unsteady two-phase flow of gas injection into stagnant water. Flow visualization data were obtained with a high-speed camera for the comparison of predicted and measured bubble sizes and shapes at various stages of the bubble growth, detachment, and gravitational rise. The CFD model is validated with these experimental measurements at different gas flow rates. The acoustic waves emitted at the time of detachment and during subsequent oscillations of the bubble were recorded with a microphone. The acoustic signature aspect of this validation is particularly interesting since it has applicability to the injection of gas into liquid mercury, which is opaque.

The Oak Ridge National Laboratory’s (ORNL) Spallation Neutron Source (SNS) facility uses a liquid mercury target that flows through a stainless steel containment vessel. As the SNS pulsed beam power level is increased, it is expected that the target vessel lifetime could become limited by cavitation damage erosion (CDE). Bubbles produced in mercury at an upwards-oriented vertical gas injector needle were observed with proton radiography (pRad) at the Los Alamos Neutron Science Center (LANSCE). The comparison of volume-of-fluid (VOF) simulation results to the radiographic images reveals some aspects of success and some deficiencies in predicting these high surface tension, highly buoyant, and non-wetting fluid behavior. Although several gas flows were measured with pRad, this paper focuses on the case with a low gas flow rate of 1.66 mg/min (10 sccm) through the 0.2-mm-outer-diameter injector needle. The acoustic waves emitted due to the detachment of the bubble and during subsequent bubble oscillations were also recorded with a microphone, providing a precise measurement of the bubble sizes. When the mercury is also motivated coaxially, the drag on the bubble forces earlier detachment leading to smaller bubble sizes.

One option that shows promise for protecting solid surfaces from cavitation damage in liquid metal spallation targets involves introducing an interstitial gas layer between the liquid metal and the containment vessel wall. Several approaches toward establishing such a protective gas layer are being investigated at the Oak Ridge National Laboratory including large bubble injection and methods that involve stabilization of the layer by surface modifications to enhance gas hold-up on the wall or by inserting a porous media. It has previously been reported that using a gas layer configuration in a test target showed an order-of-magnitude decrease in damage for an in-beam experiment. Video images that were taken of the successful gas/mercury flow configuration have been analyzed and correlated. The results show that the success was obtained under conditions where only 60% of the solid wall was covered with gas. Such a result implies that this mitigation scheme may have much more potential. Additional experiments with gas injection into water are underway. Multi-component flow simulations are also being used to provide direction for these new experiments. These simulations have been used to size the gas layer and position multiple inlet nozzles.

A combination of experimental and computational methods is necessary to adequately characterize the flow patterns in the liquid mercury target of the Spallation Neutron Source (SNS). Since liquid mercury is completely opaque and corrosive to many materials, the use of liquid mercury as the working fluid makes complete characterization of the flow field by experiment difficult. Furthermore, flow asymmetries and quasi-periodic instabilities that are observed in early target flow experiments are difficult to capture in computational fluid dynamics (CFD) simulations of the system. Therefore, an experimental program using several scaled experiments is combined with CFD simulation for the design and development of the SNS mercury target.

The Advanced Neutron Source (ANS) reactor was designed to provide a research tool with capabilities beyond those of any existing reactors. One portion of its state-of-the-art design required high-speed fluid flow through narrow channels between the fuel plates in the core. Experience with previous reactors has shown that fuel plate damage can occur when debris becomes lodged at the entrance to these channels. Such debris disrupts the fluid flow to the plate surfaces and can prevent adequate cooling of the fuel. Preliminary ANS designs addressed this issue by providing an unheated entrance length for each fuel plate so that any flow disruption would recover, thus providing adequate heat removal from the downstream, heated portions of the fuel plates. As part of the safety analysis, the adequacy of this unheated entrance length was assessed using both analytical models and experimental measurements. The Flow Blockage Test Facility (FBTF) was designed and built to conduct experiments in an environment closely matching the ANS channel geometry. The FBTF permitted careful measurements of both heat transfer and hydraulic parameters. In addition to these experimental efforts, a thin, rectangular channel was modeled using the Fluent computational fluid dynamics computer code. The numerical results were compared with the experimental data to benchmark the hydrodynamics of the model. After this comparison, the model was extended to include those elements of the safety analysis that were difficult to measure experimentally. These elements included the high wall heat flux pattern and variable fluid properties. The results were used to determine the relationship between potential blockage sizes and the unheated entrance length required.

A series of tests for static flow instability or flow excursion (FE) at conditions applicable to the proposed Advanced Neutron Source reactor was completed in parallel rectangular channels configuration with light water flowing vertically upward at very high velocities. True critical heat flux experiments under similar conditions were also conducted. The FE data reported in this study considerably extend the velocity range of data presently available worldwide. Out of the three correlations compared, the Saha and Zuber correlation had the best fit with the data. However, a modification was necessary to take into account the demonstrated dependence of the Stanton (St) and Nusselt (Nu) numbers on subcooling levels, especially in the low subcooling regime.

(2010). Distribution of Microbubble Sizes and Behavior of Large Bubbles in Mercury Flow in a Mockup Target Model of J-PARC. Journal of Nuclear Science and Technology: Vol. 47, No. 10, pp. 849-852.