Visualizing How Buildings Breathe
Two-dimensional analysis can be used for spaces where the third dimension is unimportant for the study. This is the case, for example, in a corridor where most airflow is perpendicular to the long walls.
Two-dimensional analysis could also be useful for a long space with low windows on one side and high windows on the other. But if the windows are spaced wide apart, or if the temperature fluctuation between windows is of significant interest, 3D analysis works best.
A full set of CFD analyses can be run in a few days to a week, and such simulations can continue to be tweaked along with the design—with modeling refinements made in a few hours.
Such refinements can add some cost to the analysis, especially if they involve substantial changes to the design. However, the cost of re-running the analysis with minor design changes is much lower than the cost of the first analysis, because the building model does not need to be rebuilt.
The following scenarios show a variety of conditions and the different approaches we took in each one.
Oberlin's Environmental Studies Center
A CFD simulation was recently conducted for the Adam Joseph Lewis Center for Environmental Studies at Oberlin (Ohio) College, designed by William McDonough + Partners.
This modestly scaled 13,500-square-foot (1255-square-meter) building is designed as a "net energy exporter," with highly insulated walls and roofs; energy-efficient lighting; photovoltaic panels on the roof to generate electric power; earth berming; passive solar heating and natural ventilation; and under-floor air distribution with 100 percent outside air for the heating, ventilating, and air conditioning (HVAC) system in the classroom spaces, which use heat pumps.
Each heat pump heats or cools the air in the space where it is located. The outside air is brought in via the under-floor system and is evacuated through an exhaust system. Thus, no air from one space is mechanically induced into another.
Natural ventilation was an obvious choice to scale back the size of the HVAC system, thus saving fossil-fuel-generated electricity. The 64,380-cubic-foot (1825-cubic-meter) atrium was to be naturally ventilated with high, operable windows requiring expensive motors.
CFD computer modeling determined the minimum area of operable windows that could handle the ventilation without sacrificing environmental performance and building comfort.
The modeling also determined the most efficient location for the windows. Such windows have high first costs, but a low life-cycle cost because they help eliminate the need for mechanical ventilation.
The building is oriented so that the entry to the atrium faces south into prevalent winds to maximize natural ventilation. The windows run high along the north wall and low along the east wall.
CFD studies showed that a roof-top monitor in the atrium would provide a "chimney effect" drawing fresh air in from the low windows and up through the space. But the solution was deemed too costly, primarily because of the monitor's construction and the expanse of motorized north windows.
We modeled the atrium again after reducing the number of high, motorized north windows by two-thirds. The results revealed that some uncomfortably warm areas would be created along the south wall. Based on the CFD modeling, we recommended that the designers place more operable windows low on the south wall to vent heat gain.
Rensselaer Train Station
Sometimes the goal in a multistory space is to keep the volume of air movement to a minimum, without sacrificing comfort, in order to reduce the size and power of the HVAC system, thus saving money while conserving energy.
CFD modeling of the new Rensselaer Train Station in Albany, New York, designed by Stracher-Roth-Gilmore, Architects, investigated this circumstance. This study was performed with co-funding from the New Construction Program of the New York State Energy Research and Development Authority.
This 50,000-square-foot (4,645-square-meter) building will have a double-height waiting room, crowned with a sloped roof. No natural ventilation was possible in the space because the outside air is polluted from idling diesel trains and buses and from a parking garage adjacent the site.
The CFD analysis of the waiting room was to determine how comfortable temperatures could be maintained without excessive air flow.
The model was simplified to keep the cost of the analysis down. Because it was not necessary to accurately predict temperatures at roof level, we approximated the triangular shapes of the sloped cathedral ceiling, for example, with rectangular geometry.
However, the solar radiation received by the roof was calculated from the triangular geometry, so that heat gain was accurately calculated. Iterative runs showed air moving through the space at different rates, under different temperature conditions.
The drawings shown here revealed temperature differences ranging from 105 degrees Fahrenheit (40 degrees Celsius) near the ceiling to 70 degrees F. (21 degrees C.) at floor level.
The temperature predicted is comfortable in the occupied zones, so we are now studying how much we can reduce the airflow. If the airflow rate can be significantly reduced while maintaining comfort, the HVAC designers will have two options: (1) specify smaller fans or (2) specify the same size of fans, but with variable-speed drives.
Fans with variable-speed drives can operate as constant-volume fans if the fan speed is set to yield the airflow that the CFD analysis predicts is sufficient. If the CFD prediction is borne out, this remains the final setting.
Alternately, it is possible that changes in the building's occupancy may result in higher heat gains than modeled, or perhaps the operators of the train station desire lower space temperatures during peak periods.
If this is the case, the fan speed can be reset to a higher level (but most likely still below the design level, which includes safety factors), to meet the unusual circumstances. If these circumstances are transient (e.g., some unusual grand event held in July at 4:00 p.m.), the speed can later be reset back to its initial position.
A Single Room
Although CFD was born from the need to model large-scale, complex spaces, there is no reason this high-powered simulation technology can't be used effectively to simulate small spaces, such as one room in a house.
The U.S. Department of Energy's Building America program, which funds research into making houses more energy and resource efficient, sponsored research and construction of a 1,600-square-foot (150-square-meter) prototype house in Rochester, New York.
The goal was to study how changes in a builder's routine construction techniques could lead to a more energy-efficient house. Ryan Homes, which produced more than 9,000 houses in the eastern United States in 1999, is the builder for this study. One of their objectives was to reduce the amount of HVAC ducting in order to save enough in construction costs to provide high-performance, low-emissivity (low-e) glazing.
The first experiment involved removing the heating supply registers from under the windows, a typical location in most American houses. Registers beneath windows mitigate air cooled by the surface of the windows, which otherwise flows to the floor and creates drafts.
This cool air could be eliminated with low-e glazing, which has good insulating qualities and, thus, higher radiant temperatures on the glass surface. With low-e glass, drafts are eliminated, and the supply registers can be placed in the ceilings and walls to reduce the amount of duct work.
CFD analysis confirmed that the idea works. The performance levels of the windows were improved with double-pane, low-e glazing. The simulations showed that efficient air distribution could be achieved from registers placed in the ceiling or high on the wall.
After the prototype was built, infrared thermography photos of the space verified that the CFD model was accurate. Radiant temperatures were also measured, showing that greater comfort could be achieved with the new design than with the conventional one. Ryan Homes has since incorporated these changes in their houses.
CFD computer modeling has come a long way from the time-consuming, costly simulations that were the norm only a few years ago. It is now a viable tool that can help the architect meld design ideas with building performance without sacrificing energy consumption—and can save money, too.
Adrian Tuluca, R.A., is a principal of Steven Winter Associates, Inc., an architectural research and consulting firm in Norwalk, Connecticut.