current projects

Beijing Hui Long Guan

Shenzhen Wonderland

past projects

Star Garden Site Design

     Natural ventilation
     Outdoor comfort and site planning

Beijing 12 x 12

Beijing Star Garden Initial

Beijing Star Garden Final





To design a comfortable indoor and outdoor environment requires detailed information of airflow in and around buildings. The existing simplified analyses and empirical approaches are inadequate to obtain the needed information. This paper demonstrates how the computational fluid dynamics (CFD) technique can be used by engineers to calculate the airflow distributions in and around a building. Architects can use the airflow distributions to modify their design. This design procedure shows a model of how the architects and engineers can work together to design a sustainable building.

1. Introduction

Wind is a friend of a building because the wind can naturally ventilate the building to provide a comfortable and healthy indoor environment as well as to save energy. The conventional design approach often ignores opportunities for innovation that can condition buildings at lower cost or with higher air quality and acceptable thermal comfort level, usually by means of passive cooling or natural ventilation. Figure 1 breaks the reference weather data in Beijing into different comfort categories.  Although heating is an important issue in Beijing, the results also indicate that buildings in Beijing require summer cooling. However, the figure illustrates a significant reduction of mechanical cooling if the mean air velocity in the buildings can be increased to 2 m/s by natural ventilation. Although not shown in the figure, air conditioning may not be necessary in Beijing, if natural ventilation is combined with night cooling. To design a sustainable building for Beijing, it is important to explore the opportunity to use natural ventilation.

On the other hand, the wind can be an enemy of the building because the wind can cause discomfort to pedestrians if the wind speed around the building is too high. In Beijing, there are many failing buildings, among which the wind speed on some winter days can be as high as grade 5 or 6 (8 m/s to 14 m/s). Considering the low air temperature in the winter, the chilling effect of the wind is so strong that pedestrians cannot walk comfortably and safely. Therefore, it is essential to reduce the wind speed around buildings.

For small-scale buildings, architects know something about passive solar heating, but natural ventilation and outdoor thermal comfort are very difficult to design even for a simple case. The purpose of the present study is to demonstrate, with the help of the computational fluid dynamics (CFD) technique, how architects can work with engineers to design a sustainable building.

Traditionally, an architect “predicts” the airflow in and around buildings by smart arrows as shown in Figure 2. To draw the airflow correctly requires rich knowledge of winds. Furthermore, the smart arrows cannot give the wind speed or, at least, the reliable air speed, which is an important parameter for evaluating the benefits of natural ventilation and outdoor comfort.

Figure 1 Comfort hours in Beijing

Figure 2 Smart arrows used by architects (Moore, 1993)

The engineers, on the other hand, often use a wind tunnel to simulate and measure the airflow around buildings and an environmental chamber to determine natural ventilation. Although the experimental approach provides reliable information concerning airflow in and around buildings, the available data are limited due to the expensive experimental process. Moreover, the approach is not practical for a designer to optimize the designs because the experimental method is very time consuming.

An alternative is to determine the airflow distributions in and around a building by numerical simulation. There are two widely used numerical methods. The first one is the zonal method, which calculates inter-zonal airflow through the Bernoulli equation. The prediction of the inter-zonal airflow relies on the external pressure distribution caused by wind or buoyancy effect. However, the determination of the external pressure is very complex, since the pressure distribution depends on the incoming wind speed and direction, the building size and shape, the size and location of the building interior opening (Vickery and Karakatsanis, 1987). Therefore, the accuracy of the zonal method is not good. Furthermore, the zonal model is incapable of determining thermal comfort around a building.

The other numerical method calculates the airflow distribution in and around buildings by the CFD technique. The CFD technique numerically solves a set of time-averaged partial differential equations for the conservation of mass, momentum (Navier-Stokes equations), energy, and species concentrations. The solution provides the field distribution of pressure, air velocity, temperature, and concentrations of water vapor (relative humidity) and contaminants. The method, although it has some uncertainties and requires a significant amount of computing time, has been successfully used to predict airflow in and around buildings (Chen, 1997 and Murakami, 1998). With the fast development in computer speed and capacity, the CFD technique seems to be a good approach and is therefore used in the present design.

Our design procedure is as follows. An architect initializes a design. Then an engineer uses the CFD technique to calculate the airflow in and around buildings. Based on the calculated results, the architect modifies the design. Several iterations are necessary until satisfactory indoor and outdoor environment for the building is achieved. Since an architect generally does not have sufficient fluid dynamics knowledge and numerical skills, the role of the engineer is essential to help the architect to obtain the detailed flow information. On the other hand, the engineer normally does not know how to design a building. Thus, the engineer needs the help of the architect to make the building comfortable and healthy. This paper demonstrates this procedure through the design of a small demo building in Beijing with natural ventilation (Vanke City Garden) and the analysis of a building site in Beijing for the outdoor thermal comfort and natural ventilation (Vanke Doushi Garden).

2. Natural ventilation design in Vanke City Garden

Figure 3 shows the location of the six-story demo building (20 m high) and its surroundings in Vanke City Garden, Beijing. There is a long, mid-rise building, 40 m high, in the north and low-rise buildings of six stories in the east and south. A wide street is on the west of the demo building. Thus, Vanke City Garden is a mid- and low-rise building site, in which the outdoor thermal comfort may not be a problem. So our design in this case focuses on natural ventilation.

Figure 3 The demo building (color one) and its site (Vanke City Garden)


Figure 4 Wind rose for Beijing

The design of natural ventilation in the demo building needs indoor and outdoor airflow distributions. The wind rose in Beijing (Chen et al. 1994) as shown in Figure 4 indicates that, in the summer, the prevailing wind is from north and south. The corresponding mean wind speed is 1.9 m/s.

With the wind information and the building site, a CFD program was used to calculate the airflow distribution around the demo building as shown in Figure 5. With the north wind, the air speed around the demo building is about 0.1 m/s ~ 0.4 m/s, which is very low. This is because the tall building in the north blocks the wind. The demo building is in fact in a recirculation zone of the tall building. When the wind is from the south, the air speed around the demo building is not high either (0.3~0.8 m/s). This is mainly attributed to the small distance between the demo building and those buildings in the south that block the south wind.

Nevertheless, we decided to use the available low velocity wind to design natural ventilation in the demo building. The benefits of the natural ventilation design may be limited, but it is better than nothing. Since the design of natural ventilation does not increase the building construction and operation costs, it is always worthwhile to try as long as the weather data support the use of natural ventilation. Figure 6 shows two design schemes for a typical floor of the demo building. The building layout has been designed to allow a free passage of wind. This paper demonstrates how the CFD technique is used to determine the size of the court in the south. When initializing the design, we expect the court to increase natural lighting in the building, to create a social space for the building residents, as well as to scoop the wind from the south.

Figure 7 shows the airflow distribution in and around the demo building with the wind coming from the south. Please note that the results are not extracted from those shown in Figure 5. A separate CFD calculation was done to obtain the detailed airflow distribution in and around this demo building. Without separate calculation, very fine numerical grid distribution would be needed for the demo building. The computation would require a large capacity and high-speed computer. The separation of the detailed airflow information in the demo building (Figure 7) from the airflow around Vanke City Garden (Figure 5, b) reduces the computing time significantly. Since the wind speed around the demo building is 0.3~0.8 m/s with the wind from the south as shown in Figure 5 (b), the second CFD calculation for the airflow in the demo building used a uniform south wind speed of 0.5 m/s.  The calculation shows that the building layout is good for natural ventilation because the air can flow freely through the building. However, the court size does not have a significant impact on the airflow pattern and flow rate. Furthermore, the results suggest that mechanical ventilation or stack ventilation might be needed in order to enhance the ventilation in the building. These results are very important to help us make a final decision on ventilation design.




Figure 5 The airflow distribution around the demo building (section) (a) with a north wind and (b) with a south wind.

(a)                                                                                 (b)

Figure 6 A typical floor plan of the demo building, (a) with a larger court, (b) with a smaller court.


(a) (b)

Figure 7 The air velocity distribution in and around the demo building at 1.2 m above the floor. (a) with a larger court, (b) with a smaller court.

3. Outdoor comfort and site planning in Vanke Doushi Garden

The second example is to demonstrate how the architects and engineers can work together to design a comfortable outdoor and indoor environment. People often experience an outdoor comfort problem among high-rise buildings where the airflow around the buildings can be accelerated. This study uses Vanke Doushi Garden in Beijing (Figure 8) as an example to design a comfortable outdoor environment. At the same time, natural ventilation is also considered.

To design a comfortable outdoor environment, we need to know the outdoor airflow distributions. Since it is mainly the chilling effect of the wind in the winter that causes the outdoor discomfort problem, the present design investigated the airflow distribution on a winter day.  The wind rose in Beijing (Figure 4) illustrates that, in the winter, the prevailing wind is from the north (5o inclined to the west). In a typical year, there are 9 days during which the wind speed is higher than 7.6 m/s in Beijing (ASHARE, 1997) and high wind days generally occur in the winter. The present investigation studied a scenario with a north wind of 7.6 m/s for outdoor thermal comfort consideration.

Figure 8 (a) shows the preliminary site design (Scheme I) made by an architectural firm. Design I used 16 high-rise buildings ranging from 33 to 90 m high.  This paper presents the wind speed distribution at the height of 1.5 m above the ground to evaluate pedestrians’ comfort  (Figure 9). The wind speed at section 1-1 is around 8 ~ 9m/s (Grade 5), too high to be accepted even for a short stay in the winter. The reason is that the wind can pass freely through the linear arrangement of the buildings. Furthermore, the CFD calculation shows that at the height of 30 m, the wind speed among most of the buildings is 9~10 m/s, and at the height of 70 m, the wind speed is above 12 m/s (Grade 6). The high wind speed leads to excessive high infiltration in the winter and difficulties in using the wind for natural ventilation in the summer. Therefore, the building site should be redesigned, and the height of the buildings should be reduced.

Figure 8 Three designs for Vanke Doushi Garden: (a) Original design by an architectural firm (Scheme I), (b) our first design (Scheme II), and (c) our second design (Scheme III).

Figure 9 Wind velocity distribution at the height of 1.5 m above the ground around the buildings for Scheme I with a north wind.

Since the CFD calculation shows that the height of buildings in Design I causes a serious discomfort problem, our architects designed Scheme II and Scheme III, both of which have the following features:

lower building height (the building height ranges from 20 to 60 m in Scheme II, and from 20 m to 50 m in Scheme III) to reduce winter infiltration and to provide opportunity for summer natural ventilation, without compromising the population density

protection from the north wind in the winter by using relatively high buildings in the north

Figure 10 shows that the discomfort problem is greatly reduced in Scheme II, but there are still some problems. For example, in Entrances A, B, and C, the wind speed is very high because of the linear arrangement. Staggering the entrances can easily solve this problem. Moreover, a number of issues need to be carefully examined. For instance, summer natural ventilation may not be effective in Scheme II.  As shown in Figure 8 (b), more than a half of the buildings have a long side facing east or west, such as Buildings 1-8. Since the prevailing wind in the summer is from the south in this site, the buildings with the long side facing east or west may not be able to take advantage of natural ventilation, such as cross ventilation. For example, for the buildings with the long side facing north or south (Buildings 1’-3’), the wind from the south can go through the building openings and the cross ventilation works.  In addition, the orientation is not good for passive heating design and it is difficult to shade the building from the strong solar radiation in the summer.

With the results for Scheme II, the architects in our team designed Scheme III. The low-rise buildings are now tilted 45o, thus having the long side facing south-east and north-west.

In scheme III, both outdoor thermal comfort and natural ventilation are considered. When studying the outdoor thermal comfort, the incoming wind was set to be 7.6 m/s from the north. Figure 11 shows the wind distribution for Scheme III for evaluating pedestrians’ comfort. The high-rise buildings on the north side can block the high wind from the north. As a result, the wind speed in the site is small. Although at places A and B, the wind speeds are relatively high (about 7~10 m/s). The impact on the pedestrians’ comfort is small, since they are the entrances for cars. 

For natural ventilation design, it is hard to use north wind since the high-rise buildings will block the wind. To protect cold winter wind is more important in Beijing than to use natural ventilation in the summer. However, it is feasible to use south wind for natural ventilation. The mean wind speed in the summer from the south is 1.9 m/s. With such a wind speed, Figure 12 shows that the wind speed around most of the buildings at 1.5 m above the ground is above 1.0 m/s.  The wind speed is sufficiently high for natural ventilation. The tilted building arrangement helps to introduce more wind into the site. Furthermore, the staggered arrangement prevents the front buildings from blocking winds. Therefore, Scheme III provides good outdoor thermal comfort and potential to use natural ventilation. Note that Scheme III is not our final design. The design team is studying other important issues, such as sun availability, natural lighting, and energy in buildings, etc.


Figure 10 Wind velocity distribution at the height of 1.5 m above the ground around the buildings for Scheme II with a north wind.

Figure 11 Wind velocity distribution at the height of 1.5 m above the ground around the buildings for Scheme III with a north wind (7.6 m/s)


Figure 12 Wind velocity distribution at the height of 1.5 m above the ground around the buildings for Scheme III with a south wind (1.9 m/s).
4. Conclusions

This paper shows how the engineers used the computational fluid dynamics (CFD) technique to help the architects to design natural ventilation in buildings and thermal comfort around buildings. With the help of the CFD technique, the engineers can calculate the airflow distributions in and around buildings. The results were used by the architects to modify their designs. Several iterations may be necessary to design a building with satisfactory indoor and outdoor comfort environment. This design procedure, which was undertaken by a team of architects and engineers at MIT, shows that the CFD technique is a very useful tool for building design. The cooperation between the architects and engineers is necessary in designing a good building.

Although the CFD technique has a great potential for building design, it is rather time consuming. The architects and engineers should discuss initial designs based on their experience and knowledge. This discussion will reduce the iteration significantly and can speed up the design process. The CFD technique should be used only to evaluate very few final design alternatives because the commonly used commercial CFD software still needs a long time for the calculations in order to achieve an acceptable accuracy, and the interface of the CFD software is not very user-friendly.


The research was supported by the MIT Kann-Rasmussen Program. We would like to thank Christoph Ospelt for providing the analysis result of the comfort hours in Beijing.


ASHRAE, 1997, Handbook-Fundamentals, ASHRAE, Atlanta, Ch.26.30.

Chen, Dengao, Cai, Jian, 1994, The Documents Collection for Building Design (in Chinese), Vol. 1, Chinese Construction Industry Publication Inc., Beijing, P. R. China, 196 p.

Chen, Q., 1997, “Computational fluid dynamics for HVAC: successes and failures,” ASHRAE Transactions, 103 (1), p. 178-187.

Givoni B., 1997, Climate Considerations in Building and Urban Design, Van Nostrand Reinhold Inc., New York, 464 p.

Moore, F., 1993, Environment Control Systems: Heating Cooling Lighting, McGraw-Hill, Inc., New York, 427 p.

Murakami, S., 1998, “Overview of turbulence models applied in CWE-1997,” Journal of Wind Engineering and Industrial Aerodynamics, 74-76, p. 1-24.

Vickery, B.J., and Karakatsanis, C., 1987, “External pressure distributions and induced internal ventilation flow in low-rise industrial and domestic structures,” ASHRAE Transactions, 98 (2), p. 2198-2213.

back to top

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Last modified on December 1, 2000 by
M I T   B u i l d i n g   T e c h n o l o g y   G r o u p  : : D e p a r t m e n t   o f   A r c h i t e c t u r e   R m   5 - 4 1 8
7 7   M a s s a c h u s e t t s   A v e n u e  : :  C a m b r i d g e   M A   0 2 1 3 9  : :  U S A