Assessment of pedestrian wind environment in urban planning design

Landscape and Urban Planning 140 (2015) 17–28

Contents lists available at ScienceDirect

Landscape and Urban Planning journal homepage: www.elsevier.com/locate/landurbplan

Research Paper

Assessment of pedestrian wind environment in urban planning design Xing Shi a,b,∗ , Yiyu Zhu d , Jin Duan c , Runqing Shao c , Jianguo Wang a,b a

School of Architecture, Southeast University, China Key Laboratory of Urban and Architectural Heritage Conservation, Ministry of Education, China c Research Institute of Urban Planning and Spaces, Southeast University, China d Zhejiang Academy of Building Research & Design Co., China b

h i g h l i g h t s • A space-based pedestrian wind environment assessment system can be applied in urban planning. • Tolerance factor can be incorporated in the mechanical comfort and danger model for pedestrians. • CFD is a suitable technique to assist urban planners in evaluating pedestrian wind environment.

a r t i c l e

i n f o

Article history: Received 24 October 2013 Received in revised form 25 February 2015 Accepted 31 March 2015 Keywords: Pedestrian Wind Urban planning CFD Assessment

a b s t r a c t Pedestrian wind environment is one of the urban physical environments that have a significant impact on the overall wellbeing of a city and its dwellers. It is important to be able to assess pedestrian wind environment in the urban planning design stage. Such assessment is intrinsically an in-design assessment and should be based on spatial analysis. The methodology developed considers the probabilistic nature of the assessment by determining an 80-percentile ambient wind speed as the boundary condition for wind environment simulation. The assessing criteria include mechanical comfort, safety, and amplification factor. Their threshold values are based on comparisons of existing criteria. In addition, the concept of tolerance factor, which takes into account the psychological effects of different urban spaces on people, is integrated into the assessment system. The assessment methodology can be used by urban planners in parallel with the process of designing urban physical spaces. A case study of the downtown area of an urban planning project near Lake Tai in China is conducted to demonstrate the application of the assessment methodology. Two limitations of the methodology need to be noted. First, it relies on CFD techniques to obtain the wind environment and therefore, the simulation must be conducted with considerable care and circumspection. Secondly, the methodology does not include the thermal and ventilation effects of pedestrian wind. © 2015 Elsevier B.V. All rights reserved.

1. Introduction One of the major consequences of urbanization is the increasing number of megacities all around the world, especially in East Asia, South America, and some African counties. For example, China alone has 13 cities with more than 10 million people in 2010 (City Population, n.d.) and therefore, its urbanization has been extensively studied (Cook, Gu, & Wu, 2012). While rapid urbanization offers benefits in many ways such as providing better infrastructure, higher quality of education,

∗ Corresponding author at: School of Architecture, Southeast University, #2 Sipailou, Nanjing 210096, China. Tel.: +86 15905191490; fax: +86 25 83617254. E-mail addresses: shixing [email protected] (X. Shi), [email protected] (Y. Zhu), [email protected] (J. Duan), [email protected] (R. Shao), [email protected] (J. Wang). http://dx.doi.org/10.1016/j.landurbplan.2015.03.013 0169-2046/© 2015 Elsevier B.V. All rights reserved.

generally more advanced health care, more convenient daily life, etc., it has not been immune to problems including the significant deterioration of the urban physical environment. The urban physical environment includes, but is not limited to, thermal environment, acoustical environment, lighting environment, and wind environment. These physical environments directly affect the overall wellbeing of a city and its dwellers. Examples of severe urban physical environmental problems range from those in the early industrial age (Aplin, 2012) to the latest so-called smog in Beijing, China (Coghlan, Marshall, & Slezak, 2013). 1.1. Natural wind environment and urban wind environment Research into the wind environment is important for the work of urban planners, architects, civil and structural engineers, urban and building physicists, environmental scientists, real estate

18

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

developers, city managers, and other stake holders. Although the meaning of wind environment seems to be self-explanatory, there are noteworthy differences between the pedestrian wind environment, natural wind environment, and urban wind environment that refer to the wind environment in different situations. Understanding the differences among them is the basis of this paper. First of all, wind environment, in the most generic sense, is the condition of wind in an area with spatial and temporal distributions. Theoretically and mathematically speaking, wind environment can be described by Eq. (1).

v = v(x, y, z, t).

(1)

Here, the wind velocity at a certain location and a particular time is represented by a vector v; x, y, and z are three Cartesian coordinates of the location; t denotes time. In Eq. (1), the z coordinate represents the height above the ground at which the wind environment is of interest. For the pedestrian wind environment, 1.5 m is commonly used for the z value. Note that professionals in different scientific and engineering fields concern themselves with different z values. For example, urban physicists typically study the wind environment close to the ground. Structural engineers, when analyzing the wind force on high-rise buildings, consider the wind environment much higher than 1.5 m, sometimes over 100 m, above the ground. Urban climatologists focus on the wind environment at even higher altitudes. While wind environments at different levels all have impacts on cities, the pedestrian level wind environment is most important in urban planning design. Another pair of wind environments that need to be differentiated are natural wind environment and urban wind environment. Our planet earth has had wind environment from the very first day when the atmosphere was formed, probably over four billion years ago. At that time, cities did not exist. This wind environment, without any disturbance from manmade cities, can be called natural wind environment. Urban wind environment refers to the wind environment in built-up cities. It is an altered natural wind environment by the existence of cities, mainly buildings and to a less extent other objects such as trees. To develop a pedestrian wind environment assessment system suitable for urban planning, it is obvious that the absolute magnitude of urban wind must be considered. However, it is less obvious that the difference between the natural and urban wind environments should be equally emphasized. This difference reflects how much the natural wind environment is altered by the city, thus an appropriate parameter to capture the effect of urban planning design on the natural wind environment. The focus of this study is on the pedestrian level urban wind environment. To avoid verbosity, it is referred to as pedestrian wind environment throughout the paper. 1.2. Importance of urban wind environment The urban wind environment is of great interest in many aspects of science and engineering. Structural engineers and researchers have long studied urban wind environment because wind induced lateral load is one of the primary loads a structure must resist, especially for high-rise buildings, long-span bridges, arenas, and sports stadiums. The pioneering works of Davenport (1961) and others put wind engineering onto the map of modern applied science and engineering. Another important group of professionals who are interested in urban wind environment are urban and building physicists who generally concern themselves with the pedestrian wind environment at lower levels than structural engineers. Wind effects, wind comfort, wind danger, and wind climate are the subjects of interests (Blocken & Carmeliet, 2004). Other professionals

such as climatologists also study urban wind environment, sometimes with different focuses and other times overlapping interests. Assessment of urban wind environment should start with analyzing what aspects of cities it affects? A holistic view should be taken to answer this question. Urban wind environment affects at least the following aspects of city performances. • • • • • •

Mechanical comfort of people, Thermal comfort of people, City ventilation, Natural ventilation potential for buildings, Structural wind loading on buildings, Other ecosystems.

Even if we only focus on pedestrian level urban wind environment, at least three of the above six aspects remain, i.e., mechanical comfort of people, thermal comfort of people, and city ventilation. Examination of important review works and comparative studies in this field (Blocken, Hooff, & Janssen, 2013; Koss, 2006; Ratcliff & Peterka, 1990) shows that most pedestrian wind assessment studies only address mechanical comfort, with a few exceptions considering other factors such as thermal comfort (Soligo et al., 1998; Szücs, 2013). 1.3. Assessing pedestrian wind environment from an urban planning design point of view To assess pedestrian wind environment in urban planning design, the difference between post-occupancy assessment and in-design assessment needs to be clarified. Post-occupancy assessment is defined as examination of the effectiveness for human users of occupied and designed environments (Reizenstein & Zimring, 1980). Although its definition does not exclude cities, most publications use this term on buildings. Post-occupancy assessment of buildings has been a major research field. The assessment can be conducted to a variety of physical aspects, such as thermal environment (Gou & Lau, 2013), energy consumption (Bouchlaghem, Buswell, Crippsa, & Menezesa, 2012; Hokoi, Ogura, Fu, & Rao, 2013), subjective satisfaction (Hes, Jensen, Padovani, & Raj, 2011), and green building performance in general (Fowler, Henderson, Kora, & Rauch, 2010). Similarly, post-occupancy assessment can also be done to cities despite that the word “occupancy” is usually used on buildings. The essence remains the same, that is, to assess a whole or part of a city using data that can be collected after the city is built and occupied. In-design assessment, on the other hand, intends to evaluate the performances of a building or a city in design stage. Clearly, in the design stage these performances cannot be physically measured. Instead, theoretical calculations or simulations are often used to quantify the performances based on the design. Results are obtained and evaluated to provide guidance for design adjustment and optimization. In essence, in-design assessment is based on predicting the performances of built environment. It should occur simultaneously with the design and play an important role in design evolution. Comparing post-occupancy assessment with in-design assessment, the following observations can be made. • Post-occupancy assessment is generally more reliable since it uses physical measurements to assess a building or a city. The quality of in-design assessment is highly dependent on the accuracy of performance calculation, prediction, and simulation. • In-design assessment is generally more convenient and less timeconsuming than post-occupancy assessment. It often relies on performance simulation techniques and therefore, can be completed in a relatively short period of time. On the contrary,

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

post-occupancy assessment involves field survey, installation of sensors, data collection, data analysis and therefore, needs more preparation and time. Furthermore, post-occupancy assessment is susceptible to unfavorable conditions such as limited access or inclement weather. • Last but not least, in-design assessment inherently has a much more profound impact on the overall quality of a building or a city. The earlier the assessment is done, the larger impact it has on the built environment. Post-occupancy assessment is conducted after the building or the city is built and occupied. In many cases, the problems found out of post-occupancy assessment cannot be easily remedied. Another fundamental principle for developing a methodology to assess pedestrian wind environment in urban planning design is to realize that only assessing wind conditions at one or several discrete points is not adequate. Urban planning design deals with cities or part of them with various areas, ranging from several hectares to several square kilometers. Urban planners design, analyze, and evaluate continuous urban spaces, not discrete points. Therefore, pedestrian wind environment assessment for urban planning design must also be based on spatial analysis. This issue has not been well addressed in previous research on urban wind environment assessment. Last but not least, assessing pedestrian wind environment in urban planning design is not so much of a “fail or past” test. Rather, its purpose should be to compare different designs and to identify the spaces in which the pedestrian wind environment needs to be improved. Therefore, a proper assessment system should be reasonably rigorous in terms of science and convenient for urban planners to understand and use. Otherwise, it would be difficult to serve the purpose of providing guidance for design comparison and optimization. 2. Objectives The main objective of this paper is to take the perspective of urban planners and develop a practical pedestrian wind environment assessment methodology considering mechanical comfort and danger for pedestrians. The methodology is scientifically sound and addresses some of the problems that existing assessment criteria cannot solve. In other words, a balance between scientific rigor and practical usability is maintained throughout this study to ensure that the assessment methodology developed can be applied to urban planning design. The sections below start from a short literature review, followed by methodology development. The assessment methodology is summarized using a flowchart, which can be followed step by step in an urban planning design project. Finally, a case study is presented to show the application. 3. Literature review It is not the purpose of this paper to conduct a comprehensive literature review on urban wind environment assessment. Instead, relevant literature was reviewed to analyze what knowledge is available to be the basis of this study and more importantly what is lacking when an urban planner wants to assess the pedestrian wind environment of his design. The research on pedestrian wind assessment in cities can be dated back to the 1970s. Several important wind comfort criteria were proposed. Davenport and Isyumov (1975), Hunt, Mumfort, and Poulton (1976), Lawson (1978), Lawson and Penwarden (1975), Melbourne (1978), and Penwarden (1973), Penwarden and Wise (1975) proposed their wind comfort criteria in built

19

up environment. These studies signaled the beginning of serious interest on pedestrian wind assessment in the wind engineering community. Since then, researchers from Europe, North America, and Asia have studied pedestrian wind environment assessment extensively, among which several comparison studies are worth noting. Ratcliff and Peterka (1990) compared the conclusions of five wind comfort or acceptability criteria, namely the criteria proposed in Davenport and Isyumov (1975), Hunt et al. (1976), Lawson and Penwarden (1975), Melbourne (1978), and Penwarden and Wise (1975). Another comparative study (Koss, 2006) was in the framework of the European COST Action C14, Working Group 1. It covered all five wind comfort criteria compared by Ratcliff and Peterka (1990) 16 years ago and further included the criteria implemented by four countries, namely the U.K., France, Denmark, and the Netherlands. The latest overview (Blocken et al., 2013) compared four different wind comfort criteria including those proposed in Davenport and Isyumov (1975), Melbourne (1978), and Lawson (1978). The above comparative studies on wind comfort criteria draw different conclusions. However, one conclusion appears consistently across the board, that is, the different criteria can lead to very different conclusions about the wind comfort situation in the complex urban area under study (Blocken et al., 2013). The aforementioned wind comfort criteria are essentially wind “mechanical” comfort criteria because they all look into the effects that a wind with a certain speed can have on pedestrians’ mechanical comfort, from flapping clothes to blowing people over. As previously discussed, pedestrian wind environment has multiple effects on people and cities. For instance, in addition to mechanical effect, wind chill and heat stress are also contributors to pedestrian discomfort and warrant consideration (Ratcliff & Peterka, 1990). Attempts to assess pedestrian wind environment to include effects in addition to mechanical comfort have been made. Cote, Soligo, and Williams (1992) pointed out that wind force, thermal comfort, and wind chill are the components for a comprehensive pedestrian level wind comfort criterion. A quite complex assessment system of pedestrian comfort considering both mechanical comfort and thermal comfort was developed, which requires calculation of the wind force component, the thermal component, and the wind chill component using a series of theoretical models (Soligo, Irwin, Williams, & Schuyler, 1998). Several conclusions can be drawn from this short literature review. • Pedestrian wind environment assessment has been extensively studied since the 1970s. Disparities among different wind comfort criteria have been identified. • Most of the wind comfort criteria are essentially the criteria for mechanical comfort. Studies including other comfort criteria are rare. • Almost all of the studies on pedestrian wind environment assessment tackle the problem based on discrete point analysis rather than spatial analysis. Hence, they are not able to address the real needs urban planners have in design. 4. Methodology 4.1. Urban space analysis for pedestrian wind environment assessment Urban planners design cities or part of cities with a variety of functional spaces. The word “space” is a central topic in urban theory and can be discussed philosophically (Lefebvre, 2012). However, in this paper it is limited to mean the outdoor

20

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

Table 1 A categorization of urban spaces and their descriptions.

Indoor

Outdoor

Space type

Description

Indoor space Vehicle street Pedestrian street Mixed use street City square Semi-public space Green land Waterfront space Park

Enclosed by buildings and separated from outdoors. Mainly for vehicles. Exclusively for pedestrian walking. Both people and vehicles allowed. Relatively large squares that are open to the city and serve functions such as public gathering. Usually half enclosed in a city block and half open to the city. Grass with low shrubs and sometimes trees. Usually small than a park. Rarely contains buildings. Urban spaces next to rivers, lakes, or ocean. Open to public. Good for a variety of activities, mainly leisure related. Usually larger than green land and serves more complex functions.

three dimensional physical space. Obviously, pedestrian wind environment varies spatially. Therefore, a practical pedestrian wind environment assessment for urban planning design must be spacebased. Assessment criteria based on wind conditions at randomly selected locations would not suffice. The relationship between urban wind environment and urban space needs to be explored. Categorizing urban space and linking them with pedestrian wind environment are necessary. In urban design theory, categorization of urban spaces can take different forms. From the perspective of assessing urban wind environment, a suitable categorization is shown in Table 1. The entire urban space is composed of indoor space and outdoor space. Indoor space is mainly defined by buildings and beyond the scope of this paper. Urban wind environment assessment is generally concerned with outdoor space. In Table 1, the outdoor space is divided into seven mutually exclusive space types, namely vehicle street, pedestrian street, mixed use street, waterfront space, green land, park, semi-public space, and city square. Most of these space types (Figs. 1 and 2) are self-explanatory except perhaps semipublic space. As shown in Fig. 2, semi-public spaces are usually within a city block and partially enclosed by buildings around them. These spaces have access to the city and yet they are not as open or public as city squares. 4.2. Simulation of pedestrian wind environment Assessment of the pedestrian wind environment in urban planning is an in-design assessment, not a post-occupancy assessment. Therefore, field measurement is not possible. One must be able to use the planning design and predict its pedestrian wind environment with a proper level of accuracy and certainty. Computational Fluid Dynamics (CFD) simulation is probably the most commonly used technique. CFD technique did not originate from the field of urban planning. Scientists and engineers in fields such as mechanical and aerospace engineering developed and applied it in research and design (Wilcox, 1993). However, more and more urban planners and researchers incorporate CFD into planning design for its power, speed, acceptable accuracy, visualization, and other advantages. It should be noted that application of CFD in urban planning design to simulate the pedestrian wind environment is, in itself, an active research field with ongoing studies and discussions regarding its actual ability and accuracy. Therefore, guidelines for correctly applying CFD technique in urban environment research are available (Carissimo, Franke, Hellsten, & Schlunzen, 2011). Wind tunnel testing is another important means to study urban wind environment. It is considered to be more reliable and therefore, often used to benchmark CFD models and simulation results. Compared with CFD simulations, wind tunnel testing requires specialized and often large-scale wind tunnel facilities. The cost of running a wind tunnel test for urban wind environment can be significant. Since the objective of this study is to develop a practical approach to assess pedestrian wind environment in urban planning

design, knowing the wind condition is assumed to be a prerequisite. A detailed study of the accuracy or other matters of the CFD technique when applied to simulating pedestrian wind environment is beyond the scope of this paper. It is worth noting that more and more commercially available CFD programs are at urban planners’ disposal, including general purpose CFD programs such as Star CCM+TM (CD-adapco, n.d.), FluentTM (CAE Associates, n.d), and PhoenixTM (Concentration Heat and Momentum Limited, n.d.) and others CFD programs specialized in urban air environment simulation such as StreamTM (Software CRADLE Co. Ltd., n.d.). When the CFD simulation is completed, a common way to analyze the results is to examine the wind profile in a horizontal plane close to the ground, for example, 1.5 m above. In the plane, one can define two important parameters, the spatially averaged wind speed Vavg and the maximum wind speed Vmax . Vmax is the maximum wind speed in the plane. Vavg can be defined using Eq. (2): Vavg =

8 

 n ai

V j=1 j n



.

(2)

i=1

Here, ai is the proportion of the i-th wind direction, with the north direction wind being the 1st, the northeast direction the 2nd, the east direction the 3rd, etc.; Vj is the wind speed in the j-th mesh grid; n is the total number of the mesh grid in the plane. 4.3. The probabilistic nature of pedestrian wind environment assessment Assessing pedestrian wind environment needs to take a probabilistic approach since wind environment is intrinsically stochastic. The speed and direction of wind are constantly changing. Since indesign wind environment assessment needs to predict what will happen in the future, conclusions are only valid from a probabilistic point of view. In fact, most of the wind mechanical comfort criteria take a probabilistic approach. A general form of the criteria is shown in Eq. (3): P (V ≥ Vthresh ) ≤ Paccep .

(3)

Here, V is the actual wind speed; Vthresh is the pre-defined threshold wind speed; Paccep is the acceptable probability; P is the probability of the actual wind speed exceeding the threshold wind speed. The essence of Eq. (3) is to require that the probability of the actual wind speed exceeding a pre-defined threshold value shall be less than an acceptable probability so that pedestrians’ mechanical comfort is guaranteed for most of the time. The available criteria differ on the values of Vthresh and Paccep . For instance, Lawson and Penwarden (1975) mostly used 4% for Paccep and Hunt et al. (1976) used a larger Paccep of 10%. To verify whether Eq. (3) holds true or not, field measurement is needed. A large number of measured wind speed data are gathered to evaluate the exceeding probability. However, to assess the pedestrian wind environment at the planning design stage, no field measurement can be done. One has to rely on the CFD technique

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

21

Fig. 1. Three dimensional model of a city center. The city is a newly developed one located next to Lake Tai in Jiangsu province, China.

to simulate the wind environment. Therefore, how to consider the probabilistic nature in CFD simulation becomes a key question that must be answered. The following approach offers a solution. First, using CFD simulation to obtain the pedestrian wind environment based on the planning design. The critical parameter is the ambient wind speed that is imposed as the boundary condition in the CFD model. To maintain the probabilistic nature expressed by Eq. (3), it is proposed to select an ambient wind speed Va based on a yearly exceedance probability of Paccep . In other words, in 1 year time the probability of the actual ambient wind speed exceeding Va is less than Paccep , as shown in Eq. (4). The value of Va can be determined by analyzing local meteorological data. In essence, this approach moves the consideration of the probabilistic nature from the end, i.e. comparing the field measured data with the threshold wind speed, to the beginning, i.e. determining the ambient wind speed to perform the CFD simulation. From a perspective of urban planning design, Eq. (4) is more practical to use than Eq. (3). P (V ≥ Va ) ≤ Paccep

(4)

Climatic wind data can be obtained from different sources. For example, one can log on the official website of EnergyPlus, a widely used building energy and mechanical system simulation program, and download weather data (US Department of Energy, n.d.). The weather data cover more than 2100 locations, 1042 locations in the USA, 71 locations in Canada, and more than 1000 locations in 100 other countries throughout the world. For each location, the weather data are provided in three different ASCII based formats, namely *.ddy, *.epw, and *.stat. In the *.stat file, the wind direction and speed data are given. Mean hourly wind speed data are provided for all 12 months. Fig. 3 plots the mean hourly wind speed in Nanjing, China for 12 months. Note that the x axis is the number of hourly mean wind speed data instead of time. In Fig. 3, 288 hourly mean wind speed data points are plotted. Statistical analysis was conducted to the data shown in Fig. 3. In this case, the mean wind speed was determined to be 2.2 m/s and the 80-percentile wind speed 3.5 m/s. The 80-percentile wind speed of 3.5 m/s means that the probability of the actual wind speed, as a random variable, being less than or equal to 3.5 m/s

Fig. 2. Plan view of the city center with different types of urban spaces color marked. The seven space types are vehicle street, pedestrian street, mixed use street, waterfront space, green land, park, semi-public space, and city square. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

22

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

Fig. 3. Hourly mean wind speed in Nanjing, China for 12 months. The 80-percentile wind speed is determined to be 3.5 m/s.

is 80%. Therefore, the acceptable probability Paccep in Eqs. (3) and (4) is 20%. One should note that in real boundary layer climate the ambient wind for a city is more complex. For example, the mean wind speed changes with height above the ground. Different speed profiles can be assumed to conduct CFD simulations such as the logarithmic shape (Blocken, Hooff, & Janssen, 2012).

4.4. Criteria for mechanical comfort A review of research into wind mechanical comfort criteria shows that significant differences exist among different criteria. Since the objective of this paper is to take the perspective of urban planners and develop a practical pedestrian wind environment assessment approach, conducting a detailed study to modify the existing wind mechanical comfort criteria is beyond the scope of this paper. Therefore, a set of simplified wind mechanical comfort criteria (Table 2) mainly based on the one proposed by Soligo et al. (1998) is adopted for the following reasons. The criteria are based on a review of the previous studies (Davenport & Isyumov, 1975; Hunt et al., 1976; Lawson, 1978; Lawson & Penwarden, 1975; Melbourne, 1978; Penwarden & Wise, 1975) and they are broadly consistent with them. The 80% frequency used by Soligo et al. (1998) seems to be reasonable because Penwarden and Wise (1975) selected 80% based on discussions with developers and building managers. The choice of 80%, rather than some other percentage is, to some extent, flexible in that local planning authorities may wish to raise or lower the limit slightly, based on local experience (Soligo et al., 1998; Penwarden & Wise, 1975). A minor difference is that the criteria used km/h as the unit for wind velocity and in Table 2 they are converted to m/s. Note that Table 2 assesses wind comfort using different activity categories. These activity categories can be reasonably linked with the urban space categorization previously discussed.

Table 2 Criteria for wind mechanical comfort, modified based on the criteria proposed by Soligo et al. (1998). Activity category

Threshold of mean wind velocity to feel comfortable Um (m/s)

Frequency

Sitting Standing Walking Danger

2.5 3.9 5.0 14.4

80% 80% 80% 0.1%

4.5. Criteria for danger Wind with high speed can blow pedestrians over and cause danger. Most wind comfort criteria consider this matter and require a speed limit to avoid danger. From a probability point of view, the probability Paccept in Eq. (2) for danger should be extremely small since the safety of pedestrians is at stake. Table 3 compares five criteria in terms of how they assess wind induced danger and the associated probabilities. When analyzing wind danger, the ambient wind speed used to conduct CFD simulation should not be the same as that for wind mechanical comfort. It needs to take into account the gustiness of wind, i.e., a phenomenon that the instantaneous wind speed sometimes can be significantly higher than the average wind speed over an extended period of time. In Table 3, Lawson and Penwarden (1975) used a gust factor of 2.68 to define the threshold wind speed for danger Ud . The gust factor can be defined as the ratio of the maximum 3-s running-averaged wind speed to the 10-min averaged wind speed (Nakayama, Takemi, & Nagai, 2012). Using this factor and the wind data shown in Fig. 3, Eq. (5) calculates the ambient wind speed considering gustiness: Vg = Va + g · V = 2.2 + 2.68 · 0.88 = 4.56

(5)

Here, Vg represents the gust wind speed; Va is the 80-pencentile ambient wind speed; g is the gust factor;  V is the standard deviation of the wind speed as a stochastic variable. The criteria for wind induced danger can be interpreted as the probability of the maximum wind speed in a defined space exceeding 23.7 m/s shall be no higher than 2% given that the ambient wind speed considering gustiness is 4.56 m/s. 4.6. Criteria for amplification factor It has been argued that a proper pedestrian wind environment assessment method must consider its difference from natural wind environment because this difference reflects the effect of urban

Table 3 Wind danger criteria from 5 sources. Threshold wind speed for danger, Ud (m/s)

Allowed exceeding probability, Paccept

Proposed by

15.22 23.7 20 23 14.4

0.02% 2% 10% 0.025 0.1%

Davenport and Isyumov (1975) Lawson and Penwarden (1975) Hunt et al. (1976) Melbourne (1978) Soligo et al. (1998)

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

23

planning design. Amplification factor can be used to quantify this effect, as shown in Eqs. (6)–(8). AFi =

Vu,i Vn,i

AFavg =

k 

(6)

AFi ,

(7)

i=1

AFmax = max (AFi )ki=1 .

(8)

Here, AFi represents the amplification factor at the i-th evaluating point in an urban space; Vu,i the urban wind velocity at the same point; Vn,i the natural wind velocity at the same point. In most cases, Vn,i is assumed to be equal to the ambient wind speed; AFavg the spatially averaged amplification factor; AFmax the maximum amplification factor; k the number of mesh grids in the 1.5 m high horizontal plane. This definition of amplification factor is generally consistent with that presented by Blocken and Carmeliet (2004), with a minor difference being that the wind statistics at the meteorological site was used by them instead of the natural wind velocity at the same evaluating point. Bottema (1993) split the amplification factor into two parts: a design-related contribution and a terrainrelated contribution. The underlying thought is the same, i.e., the existence of a city alters the natural wind environment and urban planning design plays a role. Similar to other criteria, the assessment of amplification factor must consider spatial variations. Therefore, when a certain type of urban space is defined, two amplification factors should be discerned, namely the spatially averaged amplification factor AFavg and the maximum amplification factor AFmax . The spatially averaged amplification factor reflects the overall effect of the urban planning design on the natural wind environment. The maximum amplification factor indicates the worst case in terms of how much the design increases the wind speed. The amplification factor can be studied using wind tunnel testing or theoretical modeling. Its magnitude depends on many variables such as observation point, geometry of buildings, open space configuration, etc. Blocken and Carmeliet (2004) reported amplification factors between 1.4 and 1.6 for a building with a through-passage. Blocken and Carmeliet (2008) studied the wind amplification factor at outdoor platforms in a high-rise apartment building and found that at some measurement points the local amplification factor can be as large as 2.0. Blocken et al. (2012) used CFD technique to study the pedestrian wind environment at the campus of Eindhoven University of Technology. The wind amplification factors found are generally less than 1.0. Based on the reported amplification factors, a threshold value of 1.0 is selected to limit the spatially averaged amplification factor and 1.5 the maximum amplification factor. In other words, for a

Fig. 4. Flowchart of pedestrian wind environment assessment in urban planning design.

defined urban space, the spatially averaged wind speed shall not exceed the natural ambient wind speed and the maximum wind speed shall not exceed that by 1.5 times. 4.7. Tolerance factor People consciously or subconsciously have different expectations for the environmental quality of urban spaces. In some spaces, the expectations tend to be higher than others. A typical example

Fig. 5. Current condition of the new city being designed.

24

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28 Table 4 Tolerance levels and tolerance factors for eight outdoor urban spaces. Space type

Tolerance level

Tolerance factor, TF

Vehicle street Pedestrian street Mixed use street City square Semi-public space Green land Waterfront space Park

High Low Middle Middle Middle Middle High Low

1.2 0.8 1.0 1.0 1.0 1.0 1.2 0.8

4.8. Assessment flowchart

Fig. 6. Eight directions of the wind and their frequencies: north – 35.3%, south – 12.8%, east – 7.1%, west – 4.9, northeast – 13.6%, southeast – 12.9%, northwest – 8.1%.

is that most of the people expect to enjoy a quiet acoustical environment in a public park and can tolerate higher noise on streets. The same phenomenon can be found on pedestrian wind environment. For instance, when pedestrians are strolling in a waterfront space they are generally more tolerable of high wind because they expect it to happen. This subjective factor should not be overlooked when developing an assessment system for pedestrian wind environment. However, very few studies, if any, address this matter. Here, we propose a simplified yet effective approach to taken tolerance into account. Tolerance levels for the eight urban space types defined in Table 1 are graded as “high”, “middle”, and “low”. A tolerance factor, denoted as TF, is assigned to each grade: 1.2 for high, 1.0 for middle, and 0.8 for low. The results are shown in Table 4. This tolerance factor can be used to adjust the comfort criteria. It is worth noting that these factors and their values are somewhat arbitrary. More research to study the psychological tolerance of pedestrians for wind environment is needed.

So far, the major components of a pedestrian wind environment assessment system for urban planning design have been discussed. The criteria can be summarized using Eqs. (9)–(12). All of the threshold values are summarized in Table 5. Note that the primary activities are identified for each space type. Um is obtained by averaging the values associated with the activity shown in Table 2. For example, if the primary activities in a space are walking and standing, the Um value for this space is equal to 4.45 m/s, which is the average of 5.0 m/s for walking and 3.9 m/s for standing. Vavg ≤ TF · Um ,

(9)

Vmax ≤ Ud ,

(10)

AFavg ≤ 1.2,

(11)

AFmax ≤ 2.0.

(12)

When an urban planner needs to perform a pedestrian wind environment assessment in a planning design project, he can follow the flowchart illustrated in Fig. 4. The procedure is explained step by step as follows. • Start from a 3 dimensional urban planning design. • Select the spaces for pedestrian wind environment assessment. Determine their types in accordance with Table 1.

Fig. 7. Seven urban spaces and their types: space 1 – a waterfront space, space 2 – a park adjacent to the waterfront space, space 3 – a public square, space 4 – a semi-public space surrounded by several high rise buildings, space 5 – a green land adjacent to a river, space 6 – a vehicle street running east-west, space 7 – a circular pedestrian street.

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

25

Table 5 Threshold values in the pedestrian wind environment assessment system. Space type

Primary activities

Threshold wind speed for mechanical comfort, Um (m/s)

Threshold wind speed for danger, Ud (m/s)

Allowed average amplification factor

Allowed maximum amplification factor

Tolerance factor

Vehicle street Pedestrian street Mixed use street City square Semi-public space Green land Waterfront space Park

Wa /Stb W/St W/St W/St/Sic W/St/Si W/St/Si W/St/Si W/St/Si

4.45 4.45 4.45 3.8 3.8 3.8 3.8 3.8

23.7 23.7 23.7 23.7 23.7 23.7 23.7 23.7

1.0

1.5

1.2 0.8 1.0 1.0 1.0 1.0 1.2 0.8

a b c

Represents walking. Represents standing. Represents sitting.

• Determine the 80-pencentile ambient wind speed Va and the gust wind speed Vg using Eq. (5). • Perform CFD simulation using Va and Vg as the boundary wind condition and for eight different wind directions. A total of 16 pedestrian wind environment profiles are obtained. • Based on the obtained pedestrian wind environments, calculate these parameters: (1) spatially averaged wind speed Vavg , (2) maximum wind speed Vmax , (3) spatially averaged amplification factor AFavg , and (4) maximum amplification factor AFmax . • Use Table 5 to determine the threshold wind speed for mechanical comfort Um and the tolerance factor corresponding to the space type. • If Eq. (9) is satisfied, go to the next step. Otherwise, revise the design. • If Eq. (10) is satisfied, go to the next step. Otherwise, revise the design. • If Eq. (11) is satisfied, go to the next step. Otherwise, revise the design. • If Eq. (12) is satisfied, go to the next step. Otherwise, revise the design. • The pedestrian wind environment of this particular space is deemed satisfying. The assessment is completed. 5. Case study The pedestrian wind environment assessment methodology was applied to a real urban planning project. It is a new city located next to Lake Tai in Jiangsu province, China (Fig. 5). The three dimensional design model is previously shown in Fig. 1. The scope of design covers a land with an area of 15 km2 , currently being farms, small villages, a few industrial facilities, and several rivers (Fig. 5). According to the weather data recorded by World Meteorological Station 583620, the frequency of the local meteorological wind is illustrated in Fig. 6. North direction wind is the most frequent wind with an occurrence probability of 35.3%. The downtown area of the new city was selected for the case study. It is located on the shore of Lake Tai and is approximately 100 ha in area. Following the procedure illustrated in Fig. 4, seven urban spaces of different types were identified (Fig. 7). The 80pencentile ambient wind speed and the gusty wind speed were determined to be 3.5 m/s and 4.56 m/s, respectively. The gust factor is 2.68 and the standard deviation is 0.8759 m/s. Vertical gradient of the ambient wind speed was not considered. The boundary conditions on the two sides and the top of the computing domain were set to be the inviscid wall condition. The normal gradients of all variables were set to be zero for the outflow boundary condition. For the ground surface, a logarithmic law for a smooth wall surface was applied. A commercial CFD software, Star CCM+TM , was used to simulate the pedestrian wind environment. Note that the CFD modeling

Fig. 8. (a) The downtown area and the computed domain with a size of 1000 m × 1000 m × 3000 m; (b) The model of the city center and the mesh grids. A total of 1.14 million mesh grids for the entire computed domain.

(Fig. 8) was conducted generally in accordance with the guidelines recommended by COST Action 732 (Carissimo et al., 2011) and by the Working Group of the Architectural Institute of Japan (Kataoka et al., 2008). Some important modeling setups are listed below. • Automatically generated polyhedral mesh was used. • A total of 1.14 million mesh grids were generated in the entire computed domain. • The turbulent model was standard k–ε model.

26

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

Fig. 9. The pedestrian wind environment at 1.5 m above the ground for space 4 under eight wind directions. The color scale in the middle represents the wind speed and is applicable to all eight directions.

• The simulation was run until the residuals leveled off. In general, the residuals are below 1E−4 and approaching 1E−5. Note that Reynolds-averaged Navier–Stokes (RANS) technique with standard k–ε model was used to compute the wind environment for its simplicity and fast speed. Other techniques such as Large Eddy Simulation (LES) and Direct Numeric Simulation (DNS) are available but typically require significantly more computing

resources (Harms, Leitl, Patnaik, & Schatzmann, 2011). Standard RANS technique in CFD is a widely used turbulence model and one of the three choices of approximate equations describing the physics of the flow in urban settings recommended by the European COST Action 732. However, standard RANS technique has limitations. One of them is that it assumes steady flow and therefore cannot fully capture the inherently unsteady plume dynamics driven by urban geometry (Harms et al., 2011).

Table 6 Parameters needed to assess the pedestrian wind environment for the 7 selected urban spaces. Space number

Space type

1 2 3 4 5 6 7

Waterfront space Park Public square Semi-public square Green land Vehicle street Pedestrian street

Parameters for assessment Vavg (m/s)

Vmax (m/s)

TF

AFavg

AFmax

1.71 1.46 2.55 0.57 1.46 1.60 1.36

4.89 6.12 6.53 6.39 5.52 5.33 5.37

1.2 0.8 1.0 1.0 1.0 1.2 0.8

0.49 0.42 0.73 0.16 0.42 0.46 0.39

1.07 1.34 1.43 1.43 1.19 1.17 1.16

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

Space 4 was selected as an example to illustrate the CFD simulation results. Fig. 9 shows the pedestrian wind environment at 1.5 m above ground under eight different wind directions. In each case, the city model is oriented the same with the direction of wind marked. The simulated results were exported into a spreadsheet for further analysis. Based on these data, the key parameters needed to perform the pedestrian wind environment assessment were calculated using the equations previously derived and they are summarized in Table 6. Using the parameters in Table 6, one can apply the criteria recommended in Eqs. (9)–(12) and follow the flowchart in Fig. 4 to perform the assessment. In this case, the pedestrian wind environments for all seven selected urban spaces passed the assessment. However, note that for spaces 3 and 4, the maximum amplification factors, both being 1.43, are quite close to the threshold of 1.5.

6. Conclusions and future work The pedestrian wind environment is one of the urban physical environmental factors affecting the overall wellbeing of a city and its dwellers. Assessment and optimization of pedestrian wind environment should be conducted in the urban planning design stage since it plays a vital role in determining the final quality of a city. Therefore, a scientifically sound and practically usable assessment methodology needs to be developed to assist urban planners in design. Pedestrian wind environment assessment in urban planning design is intrinsically an in-design assessment, not a postoccupancy one. It is also a multi-objective assessment since a comprehensive one should consider mechanical comfort, danger, thermal comfort, city ventilation, etc. Furthermore, to be applicable to planning design, the assessment methodology developed must address the problem based on spatial analysis. Assessment on discrete location points would not suffice for urban planners. Therefore, a spatial analysis method was proposed as the first step to assess the pedestrian wind environment in urban planning design. Based on the previous research assessment criteria for mechanical comfort and danger were proposed. The definitions of the amplification factor and the tolerance factor were introduced. These parameters can be used to assess whether or not a defined urban space has an acceptable pedestrian wind environment. Finally, a step-by-step flowchart was presented to guide urban planners through the assessment procedure. The CFD technique was selected to simulate the pedestrian wind environment in the case study. Although it has its own limitations, it is more efficient and less time-consuming compared with wind tunnel testing. When facing a two-dimensional wind environment diagram as shown in Fig. 9, an urban designer without special knowledge about CFD would feel difficult to interpret it, let alone evaluate pedestrian wind comfort. Given the methodology and the flowchart developed, the task becomes straightforward and much easier to handle. With more and more user-friendly commercial CFD programs are developed and made available to design offices, it is quite possible that an urban designer, with assistance of CFD specialists, can learn to conduct a solid wind environment simulation. This paper is, in no way, intended to solve every problem about assessing pedestrian wind environment in urban planning design. More research work should be done. First, thermal comfort and city ventilation are not considered in this paper and they are integral parts of a comprehensive pedestrian wind environment assessment system. Secondly, some threshold values are selected somewhat arbitrarily and they deserve further investigation. Finally, assessment is not the end; it is simply the means to identify what urban spaces have unacceptable pedestrian wind environment. How to

27

optimize the design to avoid the problems is another important research topic.

Acknowledgement The authors would like to thank the National Natural Science Foundation of China for its financial support (51138002).

References Aplin, K. (2012). Smoke emissions from industrial western Scotland in 1859 inferred from Lord Kelvin’s atmospheric electricity measurements. Atmospheric Environment, 50, 373–376. Blocken, B., & Carmeliet, J. (2004). Pedestrian wind environment around buildings: Literature review and practical examples. Journal of Thermal Envelope and Building Science, 28(2), 107–159. Blocken, B., & Carmeliet, J. (2008). Pedestrian wind conditions at outdoor platforms in a high-rise apartment building: Generic sub-configuration validation, wind comfort assessment and uncertainty issues. Wind and Structures, 11(1), 51–70. Blocken, B., Hooff, T., & Janssen, W. (2012). CFD simulation for pedestrian wind comfort and wind safety in urban areas: General decision framework and case study for the Eindhoven University campus. Environmental Modeling & Software, 30, 15–34. Blocken, B., Hooff, T., & Janssen, W. (2013). Pedestrian wind comfort around buildings: Comparison of wind comfort criteria based on whole-flow field data for a complex case study. Building and Environment, 59, 547–562. Bottema, M. (1993). Wind climate and urban geometry (Ph.D. thesis). Eindhoven: Technical University of Eindhoven. Bouchlaghem, D., Buswell, R., Crippsa, A., & Menezesa, A. (2012). Predicted vs. actual energy performance of non-domestic buildings: Using post-occupancy evaluation data to reduce the performance gap. Applied Energy, 97, 355–364. CAE Associates (n.d.). Computational Fluid Dynamics: ANSYS CFX and FLUENT CFD Software. Retrieved from https://caeai.com/ansys-software-support/ansyssoftware/computational-fluid-dynamics-ansys-cfx-and-fluent-cfd-software Carissimo, B., Franke, J., Hellsten, A., & Schlunzen, H. (2011). The COST 732 Best Practice Guideline for CFD simulation of flows in the urban environment: a summary. International Journal of Environment and Pollution, 44(1–4), 419–427. CD-adapco (n.d.). The World’s Most Comprehensive Engineering Simulation Inside a Single Integrated Package. Retrieved from http://http://www.cd-adapco.com/ products/star-ccm%C2%AE City Population (n.d.). Population Statistics for Countries, Administrative Areas, Cities and Agglomerations. Retrieved from http://www.citypopulation.de Coghlan, A., Marshall, M., & Slezak, M. (2013). China’s struggle to clear the air. New Scientists, 217(2903), 8–9. Concentration Heat and Momentum Limited (n.d.). What is PHOENICS? Retrieved from http://www.cham.co.uk/products.php Cook, I., Gu, C., & Wu, L. (2012). Progress in research on Chinese urbanization. Frontiers of Architectural Research, 1(2), 101–149. Cote, J., Soligo, M., & Williams, C. (1992). A discussion of the components for a comprehensive pedestrian level comfort criterion. Journal of Wind Engineering and Industrial Aerodynamics, 41–44, 2389–2390. Davenport, A. (1961). The application of statistical concepts to the wind loading of structures. ICE Proceedings, 19(4), 449–472. Davenport, A., & Isyumov, N. (1975). The ground level wind environment in builtup areas. In Proceedings of the fourth international conference on wind effects on buildings and structures. Heathrow: Cambridge University Press. Fowler, K., Henderson, J., Kora, A., & Rauch, E. (2010). Re-Assessing green building performance: a post occupancy evaluation of 22 GSA buildings (Technical report, Pacific Northwest National Laboratory). Richland, USA. Gou, Z., & Lau, S. (2013). Post-occupancy evaluation of the thermal environment in a green building. Facilities, 31(7/8) Harms, F., Leitl, B., Patnaik, G., & Schatzmann, M. (2011). Validating LES-based flow and dispersion models. Journal of Wind Engineering and Industrial Aerodynamics, 99, 289–295. Hes, D., Jensen, C., Padovani, R., & Raj, A. (2011). Building satisfaction – using thermal modeling to identify areas of building use focus for post occupancy evaluation. In 12th Conference of International Building Performance Simulation Association, Sydney, Australia. Hokoi, S., Ogura, D., Fu, X., & Rao, Y. (2013). Field survey on energy consumption due to hot water supply and cooking in Nanjing and Hefei, China. Frontiers of Architectural Research, 2(2), 134–146. Hunt, J., Mumfort, J., & Poulton, E. (1976). The effects of wind on people: new criteria based on wind tunnel experiments. Building and Environment, 11, 15–28. Kataoka, H., Mochida, A., Nozu, T., Shirasawa, T., Tominaga, Y., Yoshie, R., et al. (2008). AIJ guidelines for practical applications of CFD to pedestrian wind environment around buildings. Journal of Wind Engineering and Industrial Aerodynamics, 96, 1749–1761. Koss, H. (2006). On differences and similarities of applied wind comfort criteria. Journal of Wind Engineering and Industrial Aerodynamics, 94, 781–797. Lawson, T. (1978). The wind content of the built environment. Journal of Wind Engineering and Industrial Aerodynamics, 3(2–3), 93–105.

28

X. Shi et al. / Landscape and Urban Planning 140 (2015) 17–28

Lawson, T., & Penwarden, A. (1975). The effects of wind on people in the vicinity of buildings. In Proceedings of The Fourth International Conference on Wind Effects on Buildings and Structures. Heathrow: Cambridge University Press. Lefebvre, H. (2012). The production of space. Malden: Blackwell Publishing. Melbourne, W. (1978). Criteria for environmental wind conditions. Journal of Wind Engineering and Industrial Aerodynamics, 3(2–3), 241–249. Nakayama, H., Takemi, T., & Nagai, H. (2012). Large-eddy simulation of urban boundary-layer flows by generating turbulent inflows from mesoscale meteorological simulations. Atmospheric Science Letters, 13, 180–186. Penwarden, A. (1973). Acceptable wind speeds in towns. Building Science, 8(3), 259–267. Penwarden, A., & Wise, A. (1975). Wind environment around buildings (Building Research Establishment Report), London, UK. Ratcliff, M., & Peterka, J. (1990). Comparison of pedestrian wind acceptability criteria. Journal of Wind Engineering and Industrial Aerodynamics, 36(2), 791–800.

Reizenstein, J., & Zimring, C. (1980). Post-occupancy evaluation an overview. Environment and Behavior, 12(4), 429–450. Software CRADLE Co. Ltd. (n.d.). Stream. Retrieved June 24, 2013 from http://www.cfdengineering.hu/eng/stream.php Soligo, M., Irwin, P., Williams, C., & Schuyler, G. (1998). A comprehensive assessment of pedestrian comfort including thermal effects. Journal of Wind Engineering and Industrial Aerodynamics, 77–78, 753–766. Szücs, Á. (2013). Wind comfort in a public urban space – case study within Dublin Docklands. Frontiers of Architectural Research, 2(1), 50–66. US Department of Energy (n.d.). EnergyPlus Energy Simulation Software. Retrieved from http://apps1.eere.energy.gov/buildings/energyplus/weatherdata about. cfm ˜ Wilcox, D. (1993). Turbulence modeling for CFD. La Canada: DCW Industries.