Solving the World’s Toughest CFD Problems

## Debris Transport in a Nuclear Reactor Containment Building

This article, describing how FLOW-3D was used to model the performance of containment pools in nuclear facilities, was contributed by Tim Sande & Joe Tezak of Alion Science and Technology1.

In a pressurized water reactor nuclear power plant, the water circulated through the reactor core is enclosed in a primary piping system that is maintained at a pressure and temperature of roughly 2,080 psi and 585°F. Due to the high water pressure, a break in the piping would result in the generation of multiple debris types within containment. This would occur due to insulation being blown off the equipment and piping in the area surrounding the break. Typical examples of different types of debris that could be generated are shown (right).

## Emergency Core Cooling System (ECCS)

Following the pipe break, the emergency core cooling system (ECCS) would be activated. Containment sprays would be turned on to lower the containment building pressure and remove radioactive material from the atmosphere. Water would be injected into the core to remove decay heat and prevent a meltdown. This water would subsequently spill out of the pipe break. The water from both the containment spray and the decay heat removal is then pumped into containment by ECCS pumps from an outside tank. The volume of water pumped into containment via the spray and break flow collects on the containment floor and forms a pool.

## Sump Strainers and Debris

After the water from the outside tank has been depleted, the suction to the ECCS pumps would be switched over to one or more sumps in the containment building (an example of two sump strainers is shown to the left). The function of the sump(s) is to recirculate water from the containment pool to the pump suction. Each sump has a strainer system in place to prevent debris from being sucked into the ECCS pumps causing blockage or damage. However, debris that accumulates on the strainer may cause head losses that exceed the net positive suction head (NPSH) required by the pumps, causing them to fail, and preventing a safe shutdown of the plant. This is the crux of the Nuclear Regulatory Commission’s Generic Safety Issue (GSI) 191.

## FLOW-3D Applied to Evaluate Performance

FLOW-3D is used to model the containment pool and determine the quantity of debris that could arrive at the strainer(s). The pipe break, direct spray areas (regions where the spray enters the pool like rain), and runoff spray area (regions where spray water runs off floors at a higher elevation and enters the pool like a waterfall) are modeled as regions populated with mass-momentum source particles, which are assigned an appropriate flow rate and velocity. The latter depends on the freefall distance to the pool surface. The strainer areas are modeled as mass sinks, which draw the water from the containment pool.

The model is run with a free surface to identify significant water level changes (due to sump suction or choke points in the pool), and the RNG model is activated to predict turbulence in the pool. The ability of the destroyed insulation to travel through the containment pool is a function of its settling velocity (ability to travel in suspension) and tumbling velocity (ability to travel across the floor). The settling velocity correlates to the amount of kinetic energy needed in the pool to keep the piece of insulation in suspension. These settling and tumbling velocities are determined through flume and tank testing and are values calculated by the FLOW-3D model.

After the model reaches steady-state conditions, the FLOW-3D results are post-processed to determine the areas where the velocities are high enough to tumble the various debris types across the floor of the pool (shown in red) or areas where the turbulence is high enough to carry debris in suspension (shown in yellow).

The velocity vectors are then used in conjunction with the red and yellow areas to determine whether the flow would carry the debris toward the strainer(s). These areas are then compared to the initial debris distribution areas to determine the transport fractions for each type and size of debris.

## Conclusions

By combining debris transport testing with CFD modeling, the debris loads that the ECCS strainers must be able to withstand can be significantly reduced from the overly conservative values that must otherwise be assumed. CFD has also proved to be valuable in identifying containment pool water level changes, flow patterns in the vicinity of the ECCS strainer to support head loss testing, and plant design modifications.

1Alion Science and Technology is a consulting engineering company with the ITS Operation comprised of engineering professionals skilled at developing and completing diverse projects vital to power plant operations. Alion ITSO provides engineering, program management, system integration, human-systems integration, design review, testing, and analysis for nuclear, electrical and mechanical systems, as well as environmental services. Alion ITSO has developed a meticulous Quality Assurance Program, which is compliant with 10CFR50 Appendix B, 10CFR21, ASME NQA-1, ANSI N45.2 and applicable daughter standards. Alion ITSO has provided a myriad of turnkey services to customers, delivering the highest levels of satisfaction for almost 15 years.

## Fugitive Dust Emissions for Different Configurations

This article was contributed by Dhananjay Sharma, E.I., CFM, Hydraulic Modeling Engineer, AECOM.

The effect of wind on open aggregate storage piles is becoming a growing environmental problem around the world. A problem of this nature was observed at a 2.7 km2 iron-ore site. The facility receives ore via rail wagons, which are emptied by a car dumper. The ore then passes through a series of conveyors and transfer points and is transported to one of the stockpile rows. Fugitive dust emissions have been observed as a result of wind effects on the stockpiles.

Two different configurations (Options A and B) were modeled in FLOW-3D to study the impact of fugitive dust emissions. Option A has 36 stockpiles with 9 piles in 4 rows and Option B has a total of 16 stockpiles all in 1 extended row.

A 30 meter high barrier along the perimeter of the stockpiles could also be modeled in order to compare air velocities with and without the barrier. A wind velocity of 7.5 meters per second (m/s) referenced at 10 meters height was used to model both configurations. Four different wind directions were analyzed for both stockpile Options A and B.

## Physical and Numerical Modeling

### Initial Model Setup

To model the fugitive dust emissions in FLOW-3D, the air temperature was assumed to be 15°C. A single, uniform, incompressible fluid option was selected. Gravity of -9.81 m/s in the z-direction was used. The fluid was considered to be a viscous and turbulent flow. The two-equation (k-e) model was used to calculate turbulence with no surface friction for both Options A and B configurations.

### Initial Conditions

A velocity profile based on a 1/7 power law (approximately a logrithmic law-of-the-wall distribution) was assigned as an initial condition for each of the simulations. The reference velocities of most interest for the stockpile analysis are 12 and 7.5 m/s. An analysis was performed by increasing the wind speed and measuring its effect on velocities adjacent to the piles, which determined that Reynolds scaling holds for these velocities (i.e., linear relationship between incoming wind speed scaling and velocity scaling adjacent to the piles). FLOW‐3D simulations were then constructed using a velocity of 7.5 m/s only. The results from these simulations can be scaled to meet the 12 m/s condition.

A wind profile power law was used to extrapolate velocities for various heights above and below 7.5 m/s at 10 meters. This method of applying velocities at a boundary does not allow for terrain variations along the boundary The reference velocity was assigned at 10 meters above sea level for the West, Southwest and South wind directions. For the East wind direction, the velocity was assigned at the reference height of 10 meters above grade at the back (Y-max) boundary.

The wind profile power law was calculated at every meter up to 360 meters in the z-direction. Velocities were averaged over intervals equal to the mesh sizing. There were 10 height intervals at which velocity was assigned: 2, 4, 6, 8, 10, 20, 70, 181, 270 and 360 meters. After establishing a velocity profile, the values for each height interval were broken down into their X and Y components for each of the four wind directions (West, Southwest, South and East). The initial conditions were assigned at the exterior face of the mesh blocks, leaving sufficient horizontal space for the velocity profile to develop before reaching the stockpiles.

The Wind profile power law is:

$\displaystyle {{u}_{x}}={{u}_{r}}{{\left( {\frac{{{{z}_{x}}}}{{{{z}_{r}}}}} \right)}^{\propto }},where$

Ux = wind speed at height x
Ur = wind speed at reference height
Zx = height x
Zr = reference height
α = 1/7 ‐ atmospheric stability coefficient

### Geometry

Three stereo‐lithography (STL) files were created and incorporated in the model; the individual files correspond to the topography, stockyard and piles. Different STL files were generated for Options A and B.

### Meshing

The model domain was adjusted for each of the wind directions. The mesh size varies from 2.4 million to 3.3 million cells for Option A and 1.3 million cells for Option B. A cell size of 2 meters high by 4 meters long and 4 meters wide was used in the immediate vicinity of the piles in order to accurately resolve the velocities in that region.

### Boundary Conditions

Four boundary types were used in the stockpile simulations. For all wind directions, the top boundary (Z-max) was assigned as stagnation pressure. Depending on the direction of the wind, two of the sidewalls were assigned as outflow boundary conditions. The remaining two sidewalls were assigned a grid overlay boundary. Grid overlay allows the velocity from the initial conditions to enter the model. Nested blocks were used to create the desired mesh resolution and scale. At the interface between the nested blocks, a symmetry boundary condition was used. Symmetry allows information to be transferred between the blocks. Figure 1 shows the boundary condition setup for the West wind direction (y direction). For the other wind directions, a similar methodology was used for applying boundary conditions.

### Barrier

The baffle feature in FLOW-3D was used to create the wind barrier around the stockpiles. The baffles for both Options A and B were 30 meters high and were constructed in segments following the terrain. The barriers modeled are porous in nature. The porosity value of 34% (i.e., 34% open area) and corresponding velocity vs. pressure drop values were obtained from a barrier manufacturer. The FLOW‐3D model uses a baffle algorithm where porosity and associated flow losses can be specified. Baffles are infinitely thin and do not occupy any volume.

## Simulation Results

### Option A

For Option A, four different wind directions were analyzed and were simulated with and without a barrier for a wind speed of 7.5 m/s.

 Wind Direction Max Velocity w/o Barrier (m/s) Max Velocity with Barrier (m/s) Max Velocity Reduction West 13.586 11.278 17% Southwest 13.045 10.796 17% South 12.352 12.12 2% East 9.76 8.597 12%

The maximum velocities for each of the simulations and the reduction in maximum velocity between the barrier and no barrier case are presented in Table 1 above. The barrier has the least effect on maximum velocity for the South wind. The maximum velocity reduced by 2% with the addition of a barrier for Option A. The barrier had the greatest effect on velocity for the full pile case with a West or Southwest wind. The maximum velocity was reduced by 17% for both the West and Southwest winds.

### Option B

For Option B, four different wind directions were analyzed and simulated with and without a barrier for a wind speed of 7.5 m/s.

 Wind Direction Max Velocity w/o Barrier (m/s) Max Velocity w/Barrier (m/s) Max Velocity Reduction West 15.97 11.36 29% Southwest 15.14 9.21 39% South 13.4 10.1 24% East 12.78 7.15 44%

## Conclusion

The model results clearly showed that the addition of a barrier around the stockpiles would help reduce the velocities and prevent fugitive dust emissions. Though there is a cost associated with adding the barrier along the site, this option will help achieve the environmental norms by reducing dust emissions. It is evident from the model results that FLOW-3D can be used as an accurate and reliable tool for studying fugitive dust emissions. If further design changes and new options for iron-ore layouts are proposed, they can be easily modeled in FLOW-3D to determine the optimum configuration which is both cost and environmentally effective.

## Comparing HVAC System Designs

The content for this article was contributed by Alec Mercier of Tecsult Inc.

On a recent project, Tecsult’s HVAC (heating, air-conditioning and ventilation) system engineers had to consider two different configurations of a powerhouse’s air-conditioning system and were interested in demonstrating which one would provide the greatest level of comfort for the workers at the floor level. FLOW-3D was used to simulate and compare the alternative designs. The powerhouse is a large (90m long, 33m wide and 26m high), open-area building containing heat-generating equipment such as transformers, power bars, and lighting. The purpose of the air-conditioning system is to limit the maximum temperature in the building to 35ºC. Since the diffusers are located at the lower levels and the vents near the ceiling, the maximum air temperature occurs near the ceiling and the floor level is necessarily a few degrees cooler.

## Comparing Design Alternatives

The simulations provided a detailed comparison of temperatures and flow patterns throughout the powerhouse for both alternatives, and supplied Tecsult’s HVAC system engineers with valuable data to support their design choice. As part of the detailed design of the powerhouse, the engineering team proposed two different configurations of the air-conditioning system, and determined the number, size, and distribution of diffusers and vents, along with their respective flow rates, inlet/exit air velocities, and angle of projection. In both configurations, the system is comprised of three ducts blowing cool air into the powerhouse at a temperature of 25ºC and three return vents located near the ceiling. The difference between the two alternatives lies in the position of one of the air ducts. In the first alternative, the air duct runs along the upstream wall, 5.6m above the floor. In the second alternative, the air duct runs along the downstream wall, 11.1m above the floor. Figure 1 shows the powerhouse and the layout of the air ducts corresponding to alternative 1.

The most important phenomena that govern flow patterns in this type of problem are the convection currents induced by heat-generating equipment, which are usually located on the floor, and the variation in air density with temperature, i.e., buoyancy.

## Modeling Diffusers and Vents with Mass/Momentum Sources

The system engineers utilized the mass/momentum sources feature in FLOW-3D, which allows air to flow into a defined area with a specified velocity and direction. Mass/momentum sources are especially useful for simulating HVAC systems with diffusers and vents since they are often numerous and located away from the edges of the computational domain. This makes their specification difficult in a structured mesh. Mass/momentum sources can be placed arbitrarily in the domain without consideration of the proximity to mesh boundaries.

## Conclusion

The simulations provided a detailed comparison of temperatures and flow patterns throughout the powerhouse for both alternatives, and supplied Tecsult’s HVAC system engineers with valuable data to support their design choice. Figure 2 shows the temperature distribution in a plan view of the powerhouse, approximately 1m above the floor. Temperatures shown are generally between 29 and 30.5ºC, with locally higher values around heat-generating equipment. The maximum temperature is around 34ºC on the left wall.

## Architects Achieve LEED Certification in Sustainable Buildings

This article was contributed by Francisco José Lara Garachana of Simulaciones y Proyectos, SL.

LEED (Leadership in Energy and Environmental Design) is a voluntary certification system that provides third-party verification of green buildings. Participation in the voluntary LEED process demonstrates leadership, innovation, environmental stewardship and social responsibility. LEED provides building owners and operators the tools they need to immediately impact their building’s performance and bottom line, while providing healthy indoor spaces for a building’s occupants.

FLOW-3D has helped to earn the credit “IEQ- Credit 2 – Increased Ventilation” in an office building in Bogotá (Colombia). In order to receive this credit, it has to be proved that the outdoor air exceeds by 30%, the ASHRAE standards rates. In this building, the outdoor air is provided by the effects of thermal buoyancy caused by a temperature difference generated by 2 glassed chimneys on the roof heated up by solar radiation. This has to be achieved under zero wind conditions.

Initial conditions of temperature in spaces and thermal loads (or losses) were obtained from an external energy simulation software. This software considers solar radiation, thermal inertia, insulation, internal loads, insulation, glass and all other parameters that define the thermal behavior of a building. The thermal loads are included in FLOW-3D giving the final temperature distribution when the CFD simulation is done.

In the figure on the left, the 2 thermal chimneys provoke the outdoor air ingress even at zero wind conditions. Ventilation efficiency, air changes per hour, local turbulences, residence time of air are some of the variables that can be inspected in the design process. In the figure on the right, FLOW-3D gives the temperature distribution inside the building so that a thermal comfort assessment can be done in the design process. The grill sizes can be adjusted to obtain desired comfort temperatures in all spaces.

Outdoor air temperature and atmospheric pressure were set up in the boundary conditions (intake and outflow air grills). FLOW-3D did the rest obtaining the detailed air movement inside the building. The simulation was carried out until energy steady-state conditions were achieved.

The building architecture and grill dimensions had to be adjusted to obtain the correct distribution of air inside the building as well as adequate comfort temperatures. After this iteration process, the model was validated and LEED credit conditions were achieved.