Natural Ventilation Options and Performance Simulation

In its concept of a truly sustainable convention center for the city of Pittsburgh, M/E/P engineer Burt Hill Kosar Rittelmann, envisioned that natural air currents off the Allegheny River would cool the main exhibition halls—at least during favorable times of the year when the halls were not densely occupied.

By Brian Ford, Nottingham University, Nottingham, UK, Camilo Diaz, WSP Environmental Ltd. and Geoff Whittle, Simulation Technology Ltd. April 1, 2004

In its concept of a truly sustainable convention center for the city of Pittsburgh, M/E/P engineer Burt Hill Kosar Rittelmann, envisioned that natural air currents off the Allegheny River would cool the main exhibition halls—at least during favorable times of the year when the halls were not densely occupied. The question for us as the ventilation consultant was: Could such large volumes be naturally ventilated effectively? And if so, what would be the design implications and benefits?

Our experience in defining the natural ventilation design solution for major spaces in Australia’s 2000 Olympic Stadium in Sydney suggested the idea was possible. There, the scheme was based on a careful evaluation of the interaction of microclimate, occupancy patterns and the dynamic of the building fabric. A similar approach was proposed and developed for the David L. Lawrence Center.

A preliminary analysis related seasonal microclimate variation to the topographical characteristics and orientation of the site and surrounding buildings, and then overlaid these characteristics with the occupancy pattern of the building. This analysis suggested that in spring and fall natural ventilation could be considered even when the building was occupied. The prevailing wind direction, the building’s riverside location, and the catenary roof form all contributed to the feasibility of naturally driven cross ventilation of the main halls.

Defining the strategy

Two basic principles define the strategy:

1) When the wind is blowing, regardless of wind direction, pressure differences will be created around the building such that, if there are building openings on the windward and leeward sides, there will be air movement between them.

2) In still air conditions, when the internal air temperature is higher than outside—and provided there are direct connections between inside and outside at low and high levels—air movement will be created between these openings.

Other questions remained: What will the pressure differences be as the result of these forces, under different climatic conditions? How will local topography and the geometry of surrounding buildings and the convention center itself affect these pressure differences? What air-change rates are required to maintain acceptable conditions within the main halls under different patterns of occupancy? What are the required areas for the different openings around the building to achieve these air-change rates?

First, the convention center’s location, on the north side of the Allegheny, exposes it to prevailing winds, which are channelled along the direction of the river valley. This tends to put the south side of the building under higher pressure than the north side. The catenary roof rising toward the north, appropriate openings at low level on the south side and high-level openings on the north side should all encourage a naturally driven air movement through the convention halls.

However, this air movement could bypass the north side of the halls and create rather stagnant areas. Therefore, an additional supply air route was provided on the north side—from the roofs of the meeting rooms—to supply the halls at a similar level to that on the south side (see figure above). This was the basic strategy put forward for testing. Of course, at different times of the year the wind direction shifts. Also, as in most urban centers, the pattern of air movement will be highly turbulent. With the geometry and disposition of the openings proposed it is therefore necessary to investigate the impact of winds from different directions on the pattern of air movement in and around the building.

Preliminary modeling

Natural ventilation schemes were tested initially by dynamic thermal modeling and computational fluid dynamics (CFD) analysis. The geometry of the building was simplified by representing five spaces within the model. Results were encouraging but indicated the need to refine the location and area of ventilation openings to achieve better distribution of air under natural ventilation mode.

The CFD model provided a more detailed description of the geometry of the space, including roof and ventilation shafts. The exhibition booths were represented, and a finer mesh defined in the model ensures a better resolution of the flow around the booths on other internal areas. The CFD modeling focused on the set-up and tear-down periods for two different times of the year: First, the model predicted internal conditions for an outside temperature of 15°C representing a spring/autumn day, and second, when the external temperatures are 22°C to represent a summer day.

Thermal performance analysis

In determining the hours that the heating and cooling plant would have to operate, two cases were established. The base case included a full-time air-conditioning scheme. It showed about 1,200 annual hrs. of heating and about 1,000 annual hrs. of cooling required to condition the facility. By comparison, the natural ventilation case showed about 7,000 hrs. of natural ventilation could be counted on, supplemented with about 1,300 hrs. of heating, 400 hrs. of cooling and roughly 800 hrs. of mechanical ventilation. These results included mechanical heating and cooling when the building is empty between exhibitions. This obviously prevents the building from overheating in summer and underheating in winter, albeit, at a cost.

In examining the net energy supplied or removed in the two cases, testing found the base case would consume approximately 31.9 kWh per sq. meter per year for heating and 50.3 kWh per sq. meter per year for cooling. The natural ventilation case actually had a slightly higher cost for heating at 32.8 kWh per sq. meter per year, but a significantly lower cooling cost at only 16.6 kWh per sq. meter per year.

Additionally, the results showed the potential of overheating during set-up and tear-down periods, especially toward the summer. But depending on the outdoor temperature, natural ventilation can be sufficient to maintain comfortable conditions, particularly outside the peak time in summer. The period when the building is unoccupied also benefits from the possibility of controlled natural ventilation to prevent unnecessary heat buildup. The airflows vary between one-half and two air changes per hour most of the time. But there are isolated periods when airflows exceed three per hour. On average the air change rate is around one per hour.

The results demonstrated that the use of controlled natural ventilation can result in substantial energy savings, especially for mechanical cooling with savings of 67% for the conditions modeled. This is mainly due to exploiting the time when external temperatures are comfortable in Pittsburgh.

Wind tunnel tests

After obtaining generally favorable results under still-air conditions, the CFD analysis was extended to examine performance under the influence of the wind at different times of the year. This work was based on data from the draft report on wind tunnel modeling of the convention center undertaken by the Boundary Layer Wind Tunnel Laboratory of the University of Western Ontario.

Three configurations were modeled in the lab: a configuration with the proposed hotel (A) ; one without the hotel (B) ; and a final configuration incorporating later design changes without the hotel (C) . From the wind tunnel analyses, it was found that relatively small changes in pressure produced by modifying the roof (C vs. B) indicated that the approximate geometry of configuration C was an adequate representation of the building. Additionally, the effect of the hotel was limited to mildly reducing the suctions (negative pressures) while maintaining near-constant positive pressures.

Therefore, data from tests on configuration C was used for further CFD analysis. This corresponds to the configuration incorporating later design changes without the hotel.

Hall A is likely to be influenced by the exposure of the west elevation cavity vents. Its counterpart, Hall C, is likely to be particularly influenced by the exposure of the east elevation cavity vents, just as Hall B cavity vents to the west and east are protected within bridge areas. Neither are exposed on a major fa%%CBOTTMDT%%ade. Positive pressures at the cavity vents will encourage inflow and reinforce the observed flow pattern developed in the absence of wind.

Regarding the wind climate, the wind tunnel report states that compared to the annual mean, wind speeds are 5% higher in spring, 20% lower in summer, 5% lower in autumn and 15% higher in winter. It is predominantly spring and autumn when natural ventilation will be most effective, corresponding to average seasonal wind speeds nearer the annual mean. As stated above, in spring and autumn it is westerly winds that are more common.

Ventilation works

Analysis demonstrated that natural ventilation can be combined with mechanical ventilation and refrigerant-based cooling to maintain comfortable conditions within the main halls of the convention center—at appropriate times.

The flow rates achieved through the natural ventilation openings during set-up and tear-down are able to provide comfortable conditions in the summer and spring periods within the main halls when the external temperatures do not exceed 22°C. Above this, mechanical ventilation and cooling may be required.

The analysis identified the minimum inlet and outlet vent areas required, and that the roof-vent area should be at least the size of the total inlet area. The ventilation openings suggested for the west and east sides of Hall C play a very important role in the ventilation performance as they dominate the flow within the space. The proposed shaft on the south side appears to work well in bringing air from the outside and forcing a downdraft to supply fresh air into the space. While the study concentrated on Hall C, there is a close parallel between this and Halls A and B, and it was argued that the natural ventilation strategy could also be applied there.

CFD results for spring/summer days indicate that southern glazing does not appear to create downdraft problems. However, some downdraft is generated along the vertical glazing of the east fa%%CBOTTMDT%%ade, but perimeter heating would counteract this problem.

Examining the impact of wind pressures on natural ventilation, overall air volume flow rates are increased, relative to still air conditions, by winds from the west (+6%), north (+40%) and east (+61%), but are reduced by southerly winds (-12%). This implies that the conclusions reached from the thermal modeling are robust, and under most wind conditions natural ventilation is viable.

The pattern of air movement varies significantly from the still-air condition, but the internal air velocities and temperatures are generally satisfactory for all conditions modeled (outdoor temperature 15°C and 40 watts per sq. meter internal heat gain). Of course, the CFD analysis provides a snapshot of conditions at a particular time, and the internal heat gains generally are much lower than the 40 watts per sq. meter except during exhibition periods. This again implies that the natural ventilation strategy is robust to changing conditions and can be viable for significant periods of time.