Cross indicator analysis between wind energy potential and urban morphology

Accepted Manuscript Cross Indicator Analysis between Wind Energy Potential and Urban Morphology

B. Wang, L.D. Cot, L. Adolphe, S. Geoffroy, S. Sun PII:

S0960-1481(17)30562-1

DOI:

10.1016/j.renene.2017.06.057

Reference:

RENE 8924

To appear in:

Renewable Energy

Received Date:

08 February 2017

Revised Date:

23 May 2017

Accepted Date:

18 June 2017

Please cite this article as: B. Wang, L.D. Cot, L. Adolphe, S. Geoffroy, S. Sun, Cross Indicator Analysis between Wind Energy Potential and Urban Morphology, Renewable Energy (2017), doi: 10.1016/j.renene.2017.06.057

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Cross Indicator Analysis between Wind Energy Potential and Urban Morphology B. Wanga*, L.D. Cotb, L. Adolphec,d, S. Geoffroyc, S. Suna

College of Architecture and Art, North China of University of Technology, 5 Jin Yuan Zhuang Rd, 100144 Beijing, China Université de Toulouse, UPS, INSA, ISAE, ENSTIMAC, ICA (Institut Clément Ader), 3 rue Caroline Aigle, 31400 Toulouse CEDEX, France c Université de Toulouse, UPS, INSA, LMDC (Laboratoire Matériaux et Durabilité des Constructions), 135, Avenue de Rangueil, 31077 Toulouse, France d Université de Toulouse, ENSA, LRA (Laboratoire de Recherche en Architecture), 83, Rue Aristide Maillol, BP 10629, 31106 Toulouse, France a

b

Abstract

Wind energy potential in built environment can be exploitable in a proper condition. With improvement of computer science, CFD simulation with complex building configuration or city models becomes accessible. This paper’s main task is to analyse the relationship between urban wind energy potential and urban morphology through indicator studies in order to promote urban wind energy development. Six typical real urban forms in neighbourhood scale were selected from the primary wind potential evaluation with urban morphological character analysis. A set of morphological indicators that are potentially correlated with wind environment were defined. Wind energy potential over the roofs of the highest buildings in each urban tissue was evaluated. Simulation results show that the two wind energetic indicators, wind potential capacity M’ and wind potential density D' vary with different urban tissues and different wind conditions. Through the correlation analysis, we found that several urban morphological indicators have significant correlation with the two wind potential indicators. With help of these morphological indicators, we can compare much accurately and easily the output of wind potential over roof of different urban tissues, before taking the time-consuming CFD simulation for many complex urban forms.

Keywords

urban form, CFD, urban wind energy, morphological indicator, wind energetic indicator.

Nomenclature

Test point wind velocity Z Test point height from the ground 𝑈0 Reference velocity at reference height H0= 100 m 𝛼 Power law exponent in the profile of vertical wind velocity L b Cell size for the faces of the building Nf Number of inflation layers for both ground and building Tg First layer thickness from the ground rg Ground inflation transition ratio Tb First layer thickness from the building wall rb Building inflation transition ratio N Number of iteration of simulation 𝜆𝑝 Floor area ratio (FAR) 𝜆𝑎 Plot ratio (PR) 𝐻 Average building height σh Standard deviation of the building heights 𝑉𝑏 Mean building volume 𝜆𝑐 Mean aspect ratio 𝑅𝑎 Absolute Rugosity 𝑅𝑟 Relative rugosity 𝑃𝑜 Porosity (absolute) 𝑆𝜃 Sinuosity 𝑂𝑎 Actual Occlusivity 𝑂𝑣 Average Occlusivity U

Corresponding author. E- mail address: [email protected] (B. Wang) *

1

ACCEPTED MANUSCRIPT 1. Introduction 21st century is the era of Smart City. By 2050 there will be 66% of the world’s population living in cities [1]. Fast urbanisation and urban sprawl are arising increasing demand of energy and great need of efficient energy management strategy. Development of renewable energy is an important part to realise future sustainable and Smart City [2]. Wind electricity as a relatively cheap green energy has been well developed over last two decades. According to the Eurobserv'ER Database, global wind power capacity at the end of 2015 has been multiplied by 25 times from 2000 [3]. However, developing wind energy in urban environment is far from what we imagine so easy. We are searching a way to develop wind energy in the angle of urban planning. Urban morphology study is used to look for some special urban forms that can promote wind energy exploitation. Those beneficial characteristics of urban form can be evaluated in urban morphology parameters and be integrated in the system of Smart City Management. Once local energy supply is efficiently connecting to smart city energy grid [4], those exploitable wind energy resources information appeared in the urban public service system will greatly encourage people to install wind turbines around their buildings [5]. Since 1990s, many interesting projects with investigation and research on the wind energy development for built environment have been undertaken [6-12]. Facing the problems of low wind speed, high wind turbulence and important environment compact, some project results and agencies hold a rather negative point of view on the urban wind turbine application [8, 12-13]. However, as technology is developing and increasing number of trials are taken, more and more researchers have found that wind energy potential in built environment can to be exploitable in presence of wind concentration effect among buildings, high-rise buildings and new small wind turbine technologies [15-20]. In fact, urban wind turbines can be usually found installed in an open area like parks or rivers, on rooftops or an exterior corner wall of relative high buildings. Architecture integrated wind turbine is one of the positive ways to increase wind capture by modifying building form or creating a wind duct. Good examples like the case of the Bahrain World Trade Center, the Tower of Strata SE1 in London and the Pearl River Tower in Guangzhou. Urban morphology is the study of urban form, which is defined by Kevin Lynch as a part of generally identified urban area corresponding to a homogeneous area in morphological point of view [21]. As great urbanisation processing and complex urban system emerging, urban morphology becomes a popular and effective method in the domain of urban planning. Like the built form, urban form can also be determined by two forces: physical and non-physical environment [22]. In the physical environment, urban form plays an important role when facing city problems like climate change [19], energy consumption and evaluation [24, 25], renewable energy development and assessment [26, 27], traffic analysis [28-30], wind comfort [15, 31, 32], air pollution [33], etc. Results of Huang and Pham [34] showed that the thermal condition and ventilation in a neighbourhood scale could be improved by adapting the building orientation and the wind effects like effects of Venturi and canyon. Kitous et al. [35] evaluated the impact of street length and symmetry on the wind canyon effect, based on the case study of the city Ksar in Algeria. Ng et al. [31] used the method of FAD (Frontal Area Density) maps with GIS (Geographic Information System) on the Hong Kong city to analyse the relation between the urban morphology and urban ventilation situation. They found that for megacity with many highrise buildings the increase of urban form permeability of the podium layer (0-15 m) can improve the city’s ventilation. Some propositions war also given on building dispositions which can induce favourable ventilation environment in the pedestrian level. Steemers [36] presented in the project PRECis the parameters of urban forms that influent the micro-climate and energy consumption. Some of the corresponding parameters between the wind environment and urban form can be drawn in the Table 1. Edussuriya et al. [37] adopted 20 urban morphological parameters to analyse the air pollution situation in 20 neighbourhoods in Hong Kong. These urban morphological parameters have more or less impact on urban wind environment, but we chose six of them that have potential close relationship with wind energy development in an urban context. Some other related urban morphological parameters were added and described in the Section 3.1. Cionco and Ellefsen [38] used numerical simulation of micrometeorological fields to explore the influence of different urban physical morphological parameters in the urban wind flow model. Some of the analyzed parameters can be useful for our study: building density, building height, building orientation. Another reference [39] analyzed 7 different plots of the city Barcelona with urban morphology methods. They evaluated 9 morphological indices in three characteristics which could summarize the degree of compactness of urban forms compactness: density, efficiency and complexity. Table 1 Corresponding parameters between wind environment and urban morphology Wind environment 1

Wind orientation

Urban morphology Directions of streets or spaces

2

ACCEPTED MANUSCRIPT 2 3 4

Wind drag coefficient and wind pressure Pollution dispersion Average velocity near the ground

5

Turbulence

6

Ventilation produced by heat

Windward city outline Porosity of the tissue null Angle, space between buildings, vegetation rugosity Configulation orientation, street aspect ratio

The CFD (Computational Fluid Dynamics) technology is an important method to research fluid movement, such as wind circulation in a built environment. With development of computational technology these years, wind simulation in complex urban environment is becoming possible. Even though there exits still problems like turbulence model inexactitude or divergence, many numerical simulations work well and are validated by wind tunnel results. Actually, some Best Practice Guidelines (BPG) have been given for the case of wind flow in a built environment by researchers through their rich CFD simulation experiences [40-43]. In the early years, many researches have been done on the CFD simulation of wind with idealised building forms [15, 44-46]. The impact of building form and simple building group configuration on the wind environment were discussed. Considering the feasibility of wind energy exploitation and wind turbine installation in the built environment, we found most of the researches focus on the wind concentration effect [15-18, 45, 47-49] and wind energy exploitation over roof [46, 50, 51]. Therefore, for our research to develop wind energy potential in a dense built environment, we mainly focus on the zones over roofs of the buildings. Comparing to that, zones between buildings usually have less wind concentration effect, lower wind velocity and are less convenient to install a wind turbine. Apart from the research with simple building models, some more complex building configurations were tested in recent years. With CFD simulation and wind tunnel experiment, Campos-Arriaga [52] analysed the wind situation of different configuration of 7 fifteen-storey towers in the Jubilee Campus of the University of Nottingham. The distance of 15 m between buildings and an orientation of 260º are found to produce an optimal wind augmentation factors. A Darrieus wind turbine rotor design and CFD wind simulations with 5 specific sites in Copenhagen in the scale of small community (300 m×150 m) were carried out by Beller [17]. Some interesting conceptions on the wind energy development in urban environment and some practical rotor design principals were also given. Shi et al. [53] analysed the pedestrian wind environment within a scope of urban planning design. In the case study of the planning project near Lake Tai in China, a careful complex city model was simulated with CFD to evaluate the wind mechanical comfort and safety. Srifuengfung et al. [54] utilized the method of DTM (Digital Terrain Model), GIS as well as CFD to analyze the ventilation situation of 6 street blocks in the city of Bangkok. Five urban design parameters were applied to describe the urban form properties and the results found that the most important influential parameters were the urban density and the FAR (Frontal Area Rate). As for urban wind energy assessment, those study cases conducted in a neighbourhood scale or a city-scale are worthwhile for our research. Apart from some papers mentioned above with CFD [16-22, 52] some others papers used different methods to predict urban wind potential. Janajreh et al. [55] assessed energy in the Masdar City, with the help of local measured wind data and Weibull distribution. Millward-Hopkins et al. [56] used the Digital elevation models (DEMs) built upon the LiDAR (Light detection and ranging) data to construct the model of the city Leeds. Then GIS technology and regional wind climate (NOABL database) were applied to analyse the wind energy potential over roof of every building in the city. However, the LiDAR data is required high resolution raster remote sensing by survey aircraft and not available everywhere except for some UK cities. Many cities in less developed countries may not be easy to apply this method to predict urban wind potential. In this paper, we adopted the urban morphology to evaluate urban wind potential. The relationship between urban morphological indicators and wind energy indicators is going to be analyzed through several case studies. According to indication of the high-correlated morphological indicators, the urban form type with high wind potential would be pre-selected. The CFD code -- ANSYS FLUENT was used to analyze the impact of urban morphology on wind flow. Before the simulation of the real urban tissues in neighbourhood scale, a serial of simple building with different building form, some idealised urban tissues, modified urban tissue units were chosen for wind simulation and more specific research [57]. 3

ACCEPTED MANUSCRIPT 2. Methodology of validation 2.1. Parameter validation Considering the uncertainty and risks that a numerical simulation always poses, an validation from the wind tunnel experiment or in-site measure is usually indispensable. In order to make the results more plausible and reasonable, some Best Practice Guidelines (BPG) have been drawn up at least for the case of wind flow in a built environment by researchers through their validated CFD simulations results. We did a careful parameter study on the code ANSYS Fluent that we chose, with a reference tunnel experiment that had been undertaken by the Architecture Institute of Japan [58]. Through a number of trials the best-adapted CFD parameters values for the code used were found. Results of comparison between the CFD simulation and the tunnel experiment show a rather small disparity. In fact, the general average absolute error of the velocity is found 0.37 m/s for an object velocity averaged 3.05 m/s in tunnel experiment measurement (varies from 0.10 to 6.48 m/s), that means a general error of 12%. The details of the comparison and the best-adapted CFD parameters values were given by Wang et al. [59]. Since many CFD parameters are very sensitive and influent the simulation results greatly, we cannot ensure the validity of the models with great variation of geometry form and airflow environment from the validated model, whose width (W) × length (L) × height (H) = 5 m × 20 m × 20 m. Thus we chose models with similar form and size, as well as similar wind entry condition, for the following CFD simulation. Apart from the parameter validation of the software ANSYS FLUENT that we used, several verification methods were applied before and after the simulation: 1) flow consistency in the simulation field, 2) domain size analysis, 3) grid sensibility analysis and 4) avoid random errors [60].

2.2 Simulation setting As the model scale change from a building to a district or a neighborhood (urban form), the simulation setting needs to be reviewed and modified, in presence of simulation duration, precision and validation. Most of the detailed setting of CFD was the same as the best case of validation given by ref. [59]. For the simulation of idealized urban forms and the real urban forms, some necessary adjustments of the CFD setting are needed and given in the Table 2. Table 2 Additional parameters setting for urban form simulation CFD sections Geometry (half-sphere domain, measured by its radius R)

Parameter setting Idealised urban form

Tissue of Paris

Tissue of Toulouse

Tissue of Bombay

Tissue of Barcelona

400 m

750 m

750 m

750 m

750 m

General control Mesh

Typical number of tetrahedral cells Average quality of the elements Average skewness inlet conditions

Inflation layer Number of Iterations Residue

Tissue of Beijing

1 100 m

1 000 m

Relevance centre (RC): coarse, Smoothing: low 505 222

1 174 622

1 447 081

1 169 303

1 250 964

3 133 267

2 796 897

0,533

0,502

0,618

0,557

0,528

0,605

0,571

0,325

0,342

0,282

0,321

0,325

0,296

0,297

In the wind velocity estimation equation 𝑈 =

( )

Z 𝑈0 𝐻 α, 0

Launcher option

CFD soluti on

Tissue of New York

𝑈0 = 3 m/s, 𝐻0 = 10 m and 𝛼 = 0,25

Single Precision L b = 3 m, Nf = 10, Tg = 0,5 m, rg = 1,13, Tb = 0,08 m, rb = 1,25.

L b = 6 m, Nf = 8, Tg = 0,5 m, rg = 1,13, Tb = 0,1 m, rb = 1,25.

L b = 6 m, Nf = 8, Tg = 0,5 m, rg = 1,3, Tb = 0,1m, rb = 1,15.

L b = 6 m, Nf = 8, Tg = 0,5 m, rg = 1,13, Tb = 0,1 m, rb = 1,25.

L b = 6 m, Nf = 8, Tg = 0,5 m, rg = 1,3, Tb = 0,1 m, rb = 1,25.

L b = 10 m, Nf = 5, Tg = 0,5 m, rg = 1,3, Tb = 0,1 m, rb = 1,5.

3500 < N

3000 < N < 6000

Continuity < 2×10e-6, velocity < 3×10e-7

Continuity < 2×10e-5, velocity < 10e-5

L b = 6 m, N f = 8, Tg = 0,3 m, rg = 1,3, Tb = 0,1 m, rb = 1,3.

4

ACCEPTED MANUSCRIPT For every urban tissue, we choose an evaluation zone which is in neighbourhood scale, 450 m×450 m = 202500 m2, and model the existing constructions with a resolution of 3m. To make the simulation more realistic, we represent a number of buildings around the evaluation zone within a belt of 350 to 400 m. Here in the belt only the big constructions with a height more than 6 m are taken into consideration. According to the CFD BPG, the size of the simulation domain should be variable, depending on the maximum height of the building in different urban tissues: for the tissue of Parts, Toulouse, Bombay and Barcelona, where the Hmax < 35 m, we apply a domain radius of 750 m; for the tissue of New York where Hmax < 250 m, we take a domain R = 1100 m and for the tissue of Beijing where Hmax < 100 m, we take R = 1000 m. For the mesh generation, some new parametric values are taken to find a compromise between accuracy and the number of the meshes. Several tests were performed to assess the stability of the network are carried out to confirm the choice. On the base of the case of validation [59], we try to reduce the precision of simulation in order to reduce the running time as the urban form models are much more complex and bigger than the model of one building in the validation case and the idealised urban form model. In addition, the individual precision is less important than the uniformity of the simulation setting for the different urban form models, as the goal of this phase of simulation is to compare the different wind performance between the urban tissues in relative terms, rather than to get a specific result based on the local climate condition. Thus we change the level of the Relevance Center from the "medium” to “coarse” and the Smooth from the “medium” to “low”. Proved by the mesh quality analysis, some changes on the mesh size of buildings and the inflation of the boundary layer are taken, as shown in the Table 2. All the definitions of the terms are given in the guidebook of ANSYS Company.

3. Simulation of urban forms Before working with actual complex urban tissues, some idealised models were considered to analyse the relationship between the wind flow and building configuration. The authors [61] had tested the impacts of building size and courtyard aspect separately on the wind energy potential over roof at a certain heights. By changing unit building size different building density of the tissue can be evaluated, while by changing courtyard size and numbers the impact of building porosity (explained below) can be assessed. In fact, several important urban morphology parameters were selected as the indicators to signify the impact of urban form on wind energy.

3.1 Morphological indicators For the issue of wind energy evaluation, we selected only those morphological indicators that have a close relationship with the wind flow. Hence, apart from well-known urban planning indicators like Building coverage ratio (CR) and Building floor area ratio (FAR), Average building height, rugosity, Porosity and occlusivity are used for extra analysis. The definitions are given as follows: (1) Floor area ratio (FAR) Floor area ratio, also called building coverage ratio or building density, is defined as the ratio of footprint of the buildings to the overall site area. Generally, a portion of the land often remains undeveloped for other urban functions (landscape, health, leisure, etc.): ∑ 𝐴𝑖 𝑖 𝐹𝐴𝑅 = 𝜆𝑝 = (1) 𝑆 𝑡ℎ where 𝐴𝑖 is the build area of the ground floor of the 𝑖 building, 𝑆 represents the total area of the land. The building coverage ratio is an indicator to describe the building density and the land use. (2) Plot ratio (PR) Plot ratio, also called land use coefficient, is defined as the amount of construction permitted (in planning) on a land according to its size: ∑ (𝐴𝑖𝑗) 𝑖𝑗 𝑃𝑅 = 𝜆𝑎 = (2a) 𝑆 𝑡ℎ 𝑡ℎ where 𝐴𝑖𝑗 is the built area of the 𝑗 floor of the 𝑖 building and 𝑆 represents the total area of land. If the built𝑡ℎ

up area of each floor is the same for all the buildings, and the number of floors of the 𝑖 building is 𝑁𝑖, then the Equation 2a can be written as:

5

ACCEPTED MANUSCRIPT ∑ (𝐴𝑖𝑁𝑖) 𝑖

(2b) 𝑆 The land use coefficient is an indicator to describe the construction density, especially the relationship with the construction capacity of land. (3) Average building height (𝑯) The average height of all the buildings in discussion is defined by: 𝑃𝑅 = 𝜆𝑎 =

𝐻=

𝜆𝑎 ∆𝐻

𝜆𝑝 =

∑ (𝐴𝑖𝑁𝑖)∆𝐻 𝑖

∑ 𝐴𝑖 𝑖

(3)

where ∆𝐻 is the average height of a story, usually taken as 3 m. This indicator is used for the description of the absolute rugosity of the terrain. The rugosity length of the urban boundary layer is greatly influenced by the average height of the buildings. (4) Standard deviation of the building heights (𝛔𝐡) 2 σh = ∑𝑖𝑝𝑖(𝐻𝑖 ‒ 𝐻)

(4)

𝑡ℎ

where 𝑝𝑖 is the proportion of the coverage area of the 𝑖 building over the total site area, which is defined by: 𝐴𝑖 𝑡ℎ 𝑝𝑖 = ∑ 𝐴𝑖, 𝐻𝑖 is the height of the 𝑖 building. 𝑖 (5) Mean building volume [60] 𝑉𝑏 =

∑ 𝑉𝑏𝑖 𝑖

𝑁

(5)

𝑡ℎ

where 𝑉𝑏𝑖 is the volume of the 𝑖 building and 𝑁 is the total number of buildings. This indicator can be used to describe the dispersion level of the buildings with different heights. (6) Mean aspect ratio Grimmond and Oke [61] have described the effect of the building coverage in terms of the aspect ratio. Here the indicator provides extra information on the links between building surfaces and the external environment in relation to the ground surface. ∑ 𝐸𝑖 𝑖 𝜆𝑐 = (6) 𝑆 𝑡ℎ where 𝐸𝑖 is the envelop area of the 𝑖 building, which includes the surfaces of all the external walls and the roofs. This indicator not only has an important impact on the thermal exchange between the building and the environment, but also can play an important role with the surface rugosity of the buildings. (7) Absolute Rugosity As a parameter to describe the roughness of a surface to resist the free wind, absolute rugosity for a city is the average obstacle height over the whole examined area [60, 62]: 𝑅𝑎 =

∑ (𝐴𝑖𝑁𝑖)∆𝐻 𝑖

𝑆 = 𝐻 ∗ 𝜆𝑝

(7)

(8) Relative rugosity The relative rugosity is defined by the standard deviation of the building volumes [63]: 𝑅𝑟 =

∑ (𝑉𝑏𝑖 ‒ 𝑉𝑏)2 𝑖

𝑁‒1

(8)

𝑡ℎ where 𝑉𝑏𝑖 represents the volume of the 𝑖 building, 𝑉𝑏represents the average volume of the buildings and s et 𝑁 represent the number of the buildings in consideration. This indicator shows the variation level of the building volumes, which has an important influence on the wind turbulence and the wind orientation. (9) Porosity (absolute) Porosity is a ratio of the emptiness volume to the entire volume. For a city, the pores are the open spaces like the street, the plaza, the river, the lawn, etc. Here we follow the definition of the absolute porosity of an urban form by Adolphe [64]: the ratio of useful open space to the total urban volume.

6

ACCEPTED MANUSCRIPT

(∑ 𝑉 + ∑ 𝑘 𝑉 ) 𝑃𝑜 = 1 ‒

𝑖

𝑜𝑖

𝑗

𝑗 𝑐𝑗

(9)

𝑉

𝑒𝑟

𝑒𝑟

where 𝑉𝑜𝑖 is the volume of the 𝑖 open space (street, park, etc.), 𝑉𝑐𝑗 is the volume of the 𝑗 courtyard, 𝑘𝑗 is the corresponding openness coefficient of the courtyard, which can be defined according to the surface of the courtyard (𝐴𝑐): 𝑘 = 30 % if 𝐴𝑐 < 50 𝑚2. 𝑘 = 20 % if 50 𝑚2 < 𝐴𝑐 < 200 𝑚2. 𝑘 = 10 % if 200 𝑚2 < 𝐴𝑐 < 3000 𝑚2. 𝑘 = 0 if 3000 𝑚2 < 𝐴𝑐. 𝑉 is the whole canopy volume of the examined urban area, namely the controlled volume given by Balocco et Grazzini [63]: 𝑉 = (𝐻𝑚𝑎𝑥 + 10)𝑆 where 𝐻𝑚𝑎𝑥 is the height of the highest building in the study area, 10 m represents the height of wind measure height in the meteorology station. According to the definition of the relative porosity by Adolphe [62], the volume of the open space equals to the useful surface of a street section, and is calculated by the equivalent hydraulic radius. In our case, without the consideration of the wind project 𝑎𝑏

2

2

direction, the useful surface can be defined as: 𝐴𝑖 = 𝜋𝑟ℎ = 𝜋(𝑎 + 𝑏) where a and b are the width and the length of the open space on plan. (10) Occlusivity Occlusivity is a characteristic of the openness to the sky of an open space. As an important indicator of openness of a city, it is defined by:

∑ 𝑁

𝑂𝑐 =

∑𝐿

( ) 𝑖

𝑏

𝐿𝑜

𝑁

(10)

where 𝐿𝑏 is the perimeter of a building in the horizontal section and 𝐿𝑜 is the perimeter of the corresponding non-built area in the same section, 𝑁 is the number of horizontal sections [62]. As in different height of horizontal section the occlusivity of a model may vary, thus we set 𝑂𝑣 to represent the average occlusivity of all the horizontal sections and 𝑂𝑎 in images to represent the actual occlusivity for each model.

4.2 Filtration of urban forms with exploitable wind potential How to choose a right urban form to simulation and evaluate its wind potential? Firstly we search in an exhaust way the urban form types existing in the world. Then we make a filtration of the urban forms with exploitable wind potential. In the literature, there are different methods to classify urban forms, whose criteria are varied: the configuration and space structure, the buildings type (defined by the height or the function), the changes over time, the housing density, the culture and region, the location of the city (in geographical sense), the position of the urban tissue in the city and the land use function [59]. To collect as many as possible the types of existing urban form in the scale of neighbourhood that may have wind potential, we made a combination of the criteria such as the location of the city, the position of urban tissue in the city, and the land use function, to show the possible variations of the present urban forms in different physical features (Table 3).

7

ACCEPTED MANUSCRIPT Table 3 Urban form classification by land use function Location

City centre

Periphery

Description C1: ancient urban core C2: habitat of the late 19th century to the early 20th century C3: modern public areas C4: collective residence after the 2nd World War C5: current collective residence C6: individual middle-rise houses C7: individual low-rise houses C8: slum housing C9: port areas C10: large open areas P1: entertainment areas, commercial areas P2: high-tech park, university campus P3: industrial areas P4: suburbs

Due to the evolution of the city space and its multi-function, the urban forms in the same category may show a great variety among them. There are also urban spaces that are being transformed and formed a mixture of different urban forms. Thus, we cannot do the CFD simulation with all of the urban form types and need a preliminary filtration process. The selection criterion is based on the favour to assess an urban form’s capacity to capture wind energy. In this phrase, all the types of urban forms are evaluated regardless of the local cultural or climatic context. Those physical features or the characteristics of urban morphology, which can directly influence the flow of the built environment ventilation, are discussed. Then, to evaluate the wind energy capacity, the following factors should be carefully considered: 1. average wind speed, which is probably the most important factor, 2. potential surface where wind energy can be collected, 3. feasibility of the socio-economic point of view and the difficulty of installing wind turbines, 4. environmental impacts. According to the 14 categories of urban forms (Table 3) we present 17 examples of urban tissues. The description of their characteristics in terms of urban morphology is given. Base on that, the wind potential in different urban forms can be evaluated approximately and the appropriateness of the installation of wind turbines and their preferred locations can be discussed (Table 4). Generally, three groups of places were concluded in terms of the appealing to install a wind turbine: • Bad or little proposed places: ancient urban core. In this case, the tissue is compact and very rough, often slowing down the passing wind. Moreover, the relative important impacts on the social environment (noise, visual disturbance, etc...) and structure security (of the many old buildings) prevent the installation of a wind turbine. • Reserved places: individual houses, slum areas, large green areas and suburbs. For the individual houses in the city centre or in periphery, it may be feasible to install a small wind turbine on the roof or in the gardens, depending on the local wind potential in a large scale. The energy yield is often small due to the limited height of the turbine rotor. Slums and large open areas in the city centre can be selected to install an independent wind turbine with a relative high mat in order to reach a high wind speed. For the large green spaces, the visual impact of the landscape is crucial; while for the slum areas, the promotion from the government and the support from the local communities are the main points to concern. • Favourable places: modern public areas, collective residence built after the 2nd World War or contemporary, urban port areas, entertainment areas, commercial areas, industrial areas and university campus. Modern public areas, often host large office towers, have great advantages to capture wind potential because of their height and wind concentration effect. Collective residences have significant wind concentration effects around the roof edges. And in port areas, large university campus and peripheries industrial areas there are low roughness to hinder wind speed and usually exert a rather small environmental impact when installing a wind turbine. Therefore, with the analysis results shown above, we can tail off the scale of urban form choice and find more easily those most possible windy urban forms. Then, further CFD simulations will be performed to check and compare these presumed wind urban forms.

8

Table 4 Classification of urban forms and provisional feasibility evaluation of the wind turbine installation Type of urban form

Characteristics of the urban form [65]

Characteristics of the wind environment

Provisional feasibility to capture wind energy

City centre: ancient urban core

Compact, dense, small size, buildings in medium height (12 ~ 15 m) and with little variety, large roughness, few open spaces extensive, often narrow, non-geometric and without order.

Relatively stable over the roofs but below, with a lot of turbulence and low wind speed.

Poorly suited for wind turbines except in large open spaces such as squares.

Example [66]

(Toulouse, France)

City centre: habitat of the early 20th century (1)

Compact, dense, buildings in medium height (18 ~ 24m), straight and structured streets, larger than in the old urban core, geometric frame, emerging trace of the industrial age.

Similar to the old urban core but a little more turbulent, low wind speed, some wind effect like the effect of canyon.

Poorly suited for wind turbines except in large open spaces, over the external corner or along the avenues where some permanent air flow appears.

(Paris, France)

City centre: habitat of the early 20th century (2)

Compact, buildings in medium height (15 ~ 24 m), divided by district of octagonal blocks and standard size (130 ~ 140 m), each block with a big yard, orthogonal streets, wide avenues (20 m), rounded-corner street s.

Few wind in the courtyard, but good ventilation may exist in higher altitude in the avenues, little turbulence, some wind effects such as the effect of corner and the effect of canalization.

Potential places to install wind turbines on the roofs of buildings at the corners of streets.

(Barceona, Spain)

9

Type of urban form

Characteristics of the urban form [65]

Characteristics of the wind environment

Provisional feasibility to capture wind energy

City centre: modern public areas (1)

Generally central business district, very compact and dense, with many large tall buildings higher than 50 m, great roughness, no big extended open space, rather regular forms and placed in order in small rectangular lots, medium sized and well-structured streets.

Unstable, much of turbulence, average wind speed, many highly frequented wind effects like corner effect, wake effect, pyramid effect and Venturi effect.

Suitable places to install wind turbines on rooftops or near the corners of the towers, environmental impact evaluation is needed to avoid great influence on the neighboring buildings.

Example [66]

(New York, USA)

City centre: modern public areas (2)

Mainly public or private offices, dense, large blocks of considerable height (40 ~ 50 m), broad and geometric streets, regular and orderly forms, large roughness.

Rather stable and regular, little turbulence, strong seasonal winds along the avenues, average wind speed, some frequently wind effects like wake effect, Venturi effect and effect of canalization.

Suitable places to install wind turbines on the roofs of towers respecting the visual impact of facades. Wind energy assessment is demanded. Wind turbines are suggested to be installed along the avenues. (Beijing, China)

City centre: modern public areas (3)

Mostly business district, dense, many large towers, large open spaces, many different forms, large roughness, orderly and geometric broad streets.

Unstable, much turbulence, wind intensity also very varied, low speed wind, some frequently wind effects like effect of corner, wake effect, pyramid effect and Venturi effect.

Very suitable places to install wind turbines on rooftops or near the corners of the towers, wind energy assessment is proposed, the wind turbines are also proposed to be installed along the avenues. (Shanghai, China)

10

Type of urban form

Characteristics of the urban form [65]

Characteristics of the wind environment

Provisional feasibility to capture wind energy

City centre: collective residence after the 2nd World War

Dense but dispersed, buildings in large or medium size, shaped in long bars or towers, large open spaces (green areas and parking), straight streets, geometric structure, very strong traces of the industrial era.

Unstable, some turbulence, existence of strong wind, frequent wind effects like effect of corner, wake effect and the Venturi effect.

Very suitable places to install wind turbines on rooftops or near the corners of the towers. Wind energy assessment is demanded.

Example [66]

(Toulouse, France)

City centre: recent collective residence

Often built after the 1970s, dense but dispersed, buildings in large or medium size (20 ~ 50 m), shaped in long bars or towers, large green spaces, well-structured, orderly streets, humane-care environment, distinctive character.

Stable and regular, average wind speed and changing with the considered places.

Locations can be selected to install wind turbines on rooftops. Independent wind turbines are also possible but should be discreet and respect the environment.

(Guangzhou, China)

City centre: individual middle-rise houses

Usually habitat area for high and middle class, compact, dense, independent buildings in medium size plan (200 ~ 600 m²) and medium height (8 ~ 20 m), various forms of houses but generally orderly, large roughness, few extended open spaces, medium sized and relative regular streets.

Relatively stable and periodic in a large scale, little turbulence above the roofs, average wind speed.

Locations can be selected to install wind turbines on rooftops of the highest towers. Wind energy evaluation and the environmental impacts assessment is demanded.

(Mombay, India)

11

Type of urban form

Characteristics of the urban form [65]

Characteristics of the wind environment

Provisional feasibility to capture wind energy

City centre: individual lowrise houses

Low population density, dispersed, small buildings and subdivisions with private garden, building height less than 12 m, orderly and monotonous form, often rectangular geometry for economic reasons, no large open spaces, little roughness.

Stable and periodic in a large scale, little turbulence, average wind speed, a little stronger over the roofs at a height greater than 15 m than in the old city center.

Low profitability of the wind turbines due to the limited installation height and significant environmental impacts.

Example [66]

(New York, USA)

City centre: slum

Very compact and dense in plan, small and tangled houses (<300 m²), low height (2 ~ 8 m), monotone shape and style, narrow and chaotic streets, low roughness, almost no extended open spaces.

Stable and periodic in a large scale, little turbulence, very low wind speed on pedestrian level, high speed over roofs.

Suitable places to install medium or large size wind turbines which can provide energy at a large scale, an option for governments and developers.

(Mombay, India)

City centre: port areas

Low density and dispersed, buildings in different size but not tall (less than 15 m), temporary buildings and storage (silos, tanks, warehouses), green belts, little orderly forms, low roughness.

Stable and regular, average turbulence, speed of wind a little higher than in downtown.

Very suitable places to install independent wind turbines or that are integrated with architecture. Environmental impacts assessment is proposed.

(Lyon, France)

12

Type of urban form

Characteristics of the urban form [65]

Characteristics of the wind environment

Provisional feasibility to capture wind energy

City centre: large open areas

Public parks that include hills, mountains, rivers and lakes in the city, few or some small buildings, large open spaces, low roughness.

Stable and regular, existing high wind speed which depends on the considered locations.

Very suitable places to install independent wind turbines. Environmental impact assessment is demanded.

Example [66]

(New York, USA)

Periphery: entertainment areas, commercial areas

Large open spaces (parks, tennis courts, swimming pools, parking lots, fast roads), few buildings (stadiums, supermarkets, warehouses), low roughness.

Generally stable and regular in a large open space, wind speed higher than in the city center, little turbulence.

Very suitable places to install independent wind turbines or which are integrated with the architecture, few constraints on environment.

(Toulouse, France)

Periphery: high-tech park, university campus,

Built in the years 1980-90s, dispersed, large open spaces (parks, squares and sports complex), well planned and structured, distinctive character, fairly large buildings (30 ~ 100 m), medium height (10 ~ 20 m).

Stable and regular, average wind speeds in suburban areas, frequent wind effects like corner effect and Venturi effect, depending on the location and time, rather pleasant wind atmosphere created by many trees.

Very suitable places to install independent wind turbines or which are integrated with the architecture, a favorable option adapted to the nature of scientific and educational park (ecological, innovative technology). (Beijing, China)

13

Type of urban form

Characteristics of the urban form [65]

Periphery: industrial areas,

Large buildings and large open spaces, rectangular forms, temporary and factory buildings, warehouses, near the interchanges of motorways and expressways, low height (4 ~ 12 m), monotone style, low density.

Characteristics of the wind environment

Provisional feasibility to capture wind energy

Unstable, some turbulence, high wind speed, unpleasant air flow atmosphere.

Very suitable places to install wind turbines that are integrated with the architecture. Economical and practical wind turbines are proposed, few constraints on environment.

Example [66]

(Toulouse, France)

Normal periphery: suburbs

Linear forms of urbanization along the major transportation routes, sparse, unbalanced development in space, small and low-rise buildings, monotonous style, low roughness.

Generally stable and regular periodic, average wind speed, pleasant wind atmosphere.

Suitable places to install wind turbines on the rooftops of the tallest buildings or independent wind turbines in the private gardens or agriculture fields near the habitat. Wind energy evaluation and environmental impact assessment are demanded. (Toulouse, France)

14

ACCEPTED MANUSCRIPT 3.3 Wind simulation of six urban tissues Here we choose six different types of urban tissue from six different cities of the world: Paris, Toulouse, Bombay, Barcelona, New York and Beijing. In general, the tissue of Paris and tissue of Barcelona are old city centre type, mainly built in the early 20 century, compact, dense, with buildings in similar height (< 24 m), and structured street. The tissue of Toulouse is a typical urban form of collective residence built after the war: dense in construction but dispersed in land occupation, with high-rise or middle-rise buildings in form of a bar. The tissue of Bombay is a typical middle-rise individual residence: dense, few extended open spaces, medium sized and relative regular streets. The tissues of New York and Beijing are mainly consisted of high-rise buildings, but their building heights, density and open space size are different. Most of the urban tissues chosen here belong to the favourable group reserved group to develop wind potential, from the primary wind environment evaluation. Some presumed poorly performed urban tissue like the tissue of Paris is chosen for the reference and for testing the rightness of the previous presumption. A community scale of 450 m×450 m is taken for analysis (see the evaluation zone in Fig. 1). A bigger scope in radius of 350 m - 400 m is given for CFD simulation. The urban morphology information for each form case is given in the Table 5. From there we can see that, for some morphological indicators (e.g. FAR, 𝐻) there are some similar values in different urban tissues, and for some other morphological indicators (e.g. σh, 𝑉𝑏) there are quite big variance among all the six models. The indicators with different values for different urban tissues can help prove the variability and reasonability of the urban form selection, and the indicator with similar values for different urban tissues are supposed to help compare the wind potential between models. The general initial conditions are all the same for the three and the validated simulation parameters mentioned above were applied. Local climates were neglected and 8 inlet directions of every 45° in a round were used for analysing an average impact of urban form on wind potential accumulation.

Fig. 1a. Urban tissues from different cities.

15

ACCEPTED MANUSCRIPT

Fig. 1b. Urban tissues from different cities.

16

Table 5 Urban form information of each urban tissue Tissue of Paris Urban form FAR PR 𝐻

σh σh/𝐻 Hmax 𝑉𝑏

λc 𝑅𝑎 𝑅𝑟 𝑃𝑜 𝑂𝑣

Tissue of Toulouse

Tissue of Bombay

Tissue of Barcelona

Tissue of New York

Tissue of Beijing

City centre core

Collective residential area

Individual building area

City centre core

Commercial center

Collective residential area

56,21% 2.94 18.48 m 1.77 m 10 % 24 m 75 016 m2 2.41 10.37 m 85 830 m3

18.66% 1.18 20.98 m 17.90 m 85 % 50 m 37 978 m2 0.78 3.92 m 48 480 m3

22.08% 1.08 15.98 m 6.25 m 39 % 35 m 3 841 m2 1.16 3.53 m 2 803 m3

56.87% 2,94 18.06 m 5.79 m 32 % 32 m 86 665 m2 2.34 10.27 m 71 722 m3

57,94% 12,08 84.29 m 59.66 m 71 % 230 m 201 852 m2 4.50 48.86 m 143 488 m3

19,93% 3,57 56.62 m 38.02 m 67 % 93 m 69 246 m2 1.41 11.28 m 50 130 m3

32.70% 82.44%

68.26% 53.43%

78.64% 57.47%

32.04% 66.47%

42.14% 60.11%

66.51% 57.60%

𝑂𝑎

17

ACCEPTED MANUSCRIPT 3.4. Simulation result and analysis As the accessible wind-exploitation heights and areas for different urban tissues are all different, comparison of wind intensity alone (as evaluated by the wind velocity ratio U/UH) is not sufficient to describe the wind capacity of a certain urban tissue. Then, we tried to evaluate with another indicator: Total wind capacity. As we know, wind turbine power is given by the equation: 𝑃 = 𝐶𝑝 ∗ 0.5 ∗ 𝜌 ∗ 𝐴 ∗ 𝑈3 (12) where 𝐶𝑝 is the coefficient of performance, ρ the air density, A the area swept by the blades and 𝑈 the free wind velocity. Therefore, in our case, to evaluate wind energy, a simplified indicator M was defined as: 𝑛 𝑀 = ∑𝑖 = 1(𝐴𝑖 ∗ 𝑈3𝑖 𝜃) (13) where Ai represents the area of the corresponding velocity Ui. The indicator M allows to evaluate wind potential in a plane, but the plane needs to adapt to the swept area of a turbine and how many turbines can be installed in this plane. In addition, considering the difficulty to calculate the indicator M in the way proposed by [59], the paper [57] found a new simple and equivalent indicator which was proved to be pertinent to the theoretical value: 𝑀' = (𝑈3)𝑚 × 𝐴 = 𝐷' × 𝐴 (14) 3 where A is the total area of the test surface, (𝑈 )𝑚 is the area-weighted average cubic velocity which can ' be defined and imported from the code FLUENT. As the indicator of wind potential density can be defined as 𝐷 , ' 3 then we can find 𝐷 = (𝑈 )𝑚. For the wind capacity of an urban tissue, here we only consider the wind potential at a certain height over the roof of the highest buildings. According to the survey of the existing wind turbines in the market, we found that the cut-in speed of a wind turbine, which is the threshold of wind speed to arrow the turbine to produce electricity, is around 1.8-2.5 m/s for a VAWT and 3 m/s for a HAWT [8]. So we tried the wind potential evaluation with two wind speed thresholds: V > 2 m/s and V > 3 m/s. All the buildings above which at Z = 10 m the wind speed is stronger than these thresholds were taken into consideration. The detailed information of the selected building groups (classed by their heights) is given by the Table 6. Some of the simulation results in four wind inlet directions (𝜃 =‒ 30°,60°, 150°, ‒ 120°) are given in Fig. 2, showing the wind velocity contour line of the six urban tissues at 10 m above the ground. The wind potentials in M’ on Z = 10 m above the roof of the highest buildings of the six urban tissues in 8 inlet wind directions were shown in Fig. 3. Table 6 ' ' Sum of wind potential capacity 𝑀 and wind potential density 𝐷 for the highest buildings in different urban tissues

Buildings with the exploitable wind energy

Paris

Toulouse

Bombay

Barcelona

New York

Heights (m)

16, 18, 20, 24

27, 40, 50

24, 35

18, 20, 24, 27, 32

80-230

Total roof surface (m2)

111346

16724

13223

88212

46174

24220

2,80

1,20

0,52

1,79

3,37

1,54

3,55

1,25

0,59

2,52

3,42

1,61

25

72

39

20

73

64

32

75

45

29

74

66

M' (106 m5s-3) with the exploitable wind energy (U > 3 m/s) M' (106 m5s-3) with the exploitable wind energy (U > 2 m/s)

𝐷' (m3s-3) with the exploitable wind energy (U > 3 m/s)

𝐷' (m3s-3) with the exploitable wind energy (U > 2 m/s)

Beijing 70, 87, 93

From the Table 6, it can be seen that: 1. The total wind potential capacity (M’) of different urban tissue varies. Among the six analyzed, the tissues of New York and Paris possess generally higher wind potential than other tissues, while the tissues of Toulouse and Bombay show the least wind potential. ' 2. The total wind potential density (𝐷 ) of different urban tissue varies. Among the six, the tissues of New York and Toulouse possess generally higher wind potential than other tissues, while the tissues of Paris and Barcelona show the least wind potential density.

18

ACCEPTED MANUSCRIPT Thus, we can see that the wind potential capacity in a given site scale is greatly influenced by the exploitable roof area of the urban tissue. High wind density with small exploitable roof area as the tissue of Toulouse shows low wind capacity, while the case of Paris is just reverse.

Fig. 2. Wind velocity contour lines of the urban tissue models in four wind inlet directions.

M' (m5*s-3)

5000000 Paris

4000000

Toulouse

3000000

Bombay Barcelona

2000000

New York

1000000

Beijing

0 15°

60°

105° Wind inlet150° direction-165° Ɵ

-120°

-75°

-30°

Fig. 3a. Comparison of wind potential capacity for different urban tissues (when the wind speed U > 2 m/s).

19

ACCEPTED MANUSCRIPT

M' (m5*s-3)

5000000 Paris

4000000

Toulouse

3000000

Bombay Barcelona

2000000

New York

1000000

Beijing

0 15°

60°

105° Wind inlet150° direction-165° Ɵ

-120°

-75°

-30°

Fig. 3b. Comparison of wind potential capacity for different urban tissues (when the wind speed U > 3 m/s).

From the Fig. 3, it can be seen that: 1. The wind performance varies from different inlet direction. For the tissues of Toulouse and Bombay, the differences are very small. As for the other four tissues, the variance of wind performance over different inlet wind direction is evident, especially the tissues of Paris and New York. Besides, those best wind performance inlet direction can be found for each tissue: Paris -30°, Toulouse -120°, Bombay -150°, Barcelona -165°, New York 60° and Beijing 15°. 2. Generally, the values M’ of tissues of Paris and New York are the highest among the six urban tissues. However, as the inlet wind direction is specified, the outcome may vary: for example, for an inlet wind direction of 60° the tissue of New York is more windy, while for a wind of -120° the tissue of Paris has more wind potential. 3. When changing the thresholds of wind speed, the variance of wind potential capacity of different urban tissue is also different: on average, New York 2%, Toulouse 4%, Beijing 7%, Bombay 15%, Paris 31% and Barcelona 43%. These values reveal to a certain degree the proportion of wind velocity interval in 2-3 m/s: the bigger the variance is, the bigger the proportion of the wind velocity in this interval is.

4. Discussion 4.1. Correlation between morphological parameters and wind potential indicators To evaluate the impact of urban morphology on the wind potential, we can analysis the correlation between different urban morphological parameters and wind potential capacity and density (on the condition of two wind speed thresholds). The Spearman correlation analysis method was used to calculate the correlation coefficient r. 𝑛

∑(𝑋 ‒ 𝑋)(𝑌 ‒ 𝑌) 𝑖

𝑟=

𝑖

1 𝑛

(15)

𝑛

∑(𝑋 ‒ 𝑋) ∑(𝑌 ‒ 𝑌) 2

𝑖

1

𝑖

2

1

Where the 𝑋𝑖 and 𝑌𝑖 represent the two groups of data in different influence aspect, 𝑋 and 𝑌 represent their average value, and 𝑛 represents the number of sample (the sample quantities are the same for the two groups of data). According to the results of Table 5 and 6, the correlation coefficients between the four wind potential indicator values and 12 morphological parameters were given as in the following Table 7. The positive value of r means positive correlation while the negative value means negative correlation. The absolute value of r varies from 0 to 1. 0 means no correlation while 1 means full correlation. To evaluate the justices of the correlation, significance test was done. The statistical magnitude t can be calculated according to the Equation 16 and the results were given in Table 8. 𝑡=

𝑟 𝑛‒2 1 ‒ 𝑟2

(16)

20

ACCEPTED MANUSCRIPT With a confidence interval of 95% and a degree of freedom of n - 2 = 4, we can get the statistical magnitude according to the t distribution function: 𝑡0.025, 4 = 2.776. In the same way, with a confidence interval of 99% the statistical magnitude 𝑡0.01, 4 = 3.747. Therefore, when t > 2.776 or t < -2.776, it means there is a significant correlation relation; while the t > 3.747 or t < -3.747 means a very significant correlation. Then the thresholds of significant correlation and very significant correlation for r are: 0.811 and 0.882. Therefore, in the Table 8, we marked the t values with significant correlation (in bold blue) and the t values with very significant correlation (in bold red). The most significant correlation relations were found between λc and M’, Rr and M’, σh/𝐻 and D'. Some other parameters show different correlation level between the two wind conditions: FAR and Po show significant correlation with M’ and U > 2 m/s; 𝑉𝑏 show significant correlation with M’ and U > 3 m/s. To evaluate the overall correlation relation between every wind potential indicator with different morphological parameter, an average value of r is calculated according to the Fisher Z-Transformation:

( )

1 1+𝑟 𝑍𝑟 = ln 2 1‒𝑟

(17)

The results are shown in the Table 7. Then we found that, compared to the wind potential density D', the capacity M’ (in both two wind threshold conditions) has relevantly higher correlation with different morphological parameters on average. Between the two wind thresholds conditions, with wind U > 3m/s there were higher correction coefficient on average. Table 7 Correlation between wind potential indicators and urban morphological parameters M' with wind M' with wind D with wind D with wind U > 3 m/s U > 2 m/s U > 3 m/s U > 2 m/s FAR 0.81 0.90 -0.44 -0.45 PR 0.80 0.64 0.42 0.40 0.59 0.37 0.67 0.65 𝐻 σh 0.48 0.24 0.79 0.78 σh/𝐻 -0.12 -0.35 0.94 0.95 Hmax 0.61 0.39 0.67 0.65 𝑉𝑏 0.89 0.77 0.31 0.29 λc 0.89 0.82 0.07 0.05 𝑅𝑎 0.79 0.63 0.42 0.41 𝑅𝑟 0.96 0.89 0.22 0.20

𝑃𝑜 𝑂𝑟 Average

-0.78 0.53 0.80

-0.91 0.71 0.75

0.47 -0.71 0.65

0.48 -0.72 0.66

Table 8 Significance test value t of the correlation coefficients between wind potential indicators and urban morphological parameters M' with wind M' with wind D' with wind D' with wind U > 3 m/s U > 2 m/s U > 3 m/s U > 2 m/s FAR 2.76 4.22 -0.98 -1.01 PR 2.66 1.65 0.94 0.89 1.47 0.81 1.80 1.70 𝐻 σh 1.10 0.50 2.61 2.47 σh/𝐻 -0.25 -0.75 5.31 5.87 Hmax 1.55 0.85 1.80 1.72 𝑉𝑏 3.81 2.44 0.65 0.61 λc 3.85 2.83 0.15 0.11 𝑅𝑎 2.60 1.62 0.94 0.89 𝑅𝑟 7.09 3.92 0.44 0.41

𝑃𝑜 𝑂𝑟

-2.50 1.25

-4.47 2.03

1.06 -1.99

1.09 -2.08

21

ACCEPTED MANUSCRIPT

Correlation coefficient r

With the same method, we made the correlation between different urban morphological parameters and wind potential capacity M’ values in different inlet wind directions (with wind speed thresholds U > 2 m/s). The correlation coefficient r values are given in Figure 4. We can see that: 1. The most correlated morphological parameters with the average wind inlet directions are: FAR, 𝑅𝑟 and 𝑃𝑜, which is the same result with those in Table 7 ; 2. The most correlated morphological parameters are also those have smallest variance between different wind inlet directions. That means these morphological parameters correlated with wind potential capacity get little influence from inlet wind direction. 3. According to the conformity of the curve shape of the correlation between morphological parameters and wind inlet directions, we find many of these parameters have great similarity, especially between PR and Ra, λc and 𝐻, Therefore, to avoid repetition, it’s better not to evaluate both two parameters in the future. 1

FAR PR H

0.5

σ_h σ_h/H Hmax

0 15°

60°

105°

150°

-165°

-120°

-75°

-30°

(Vb ) λ_c

-0.5

Ra Rr

-1

Wind inlet direction Ɵ

Po Or

Fig. 4 Correlation between wind potential in different inlet wind directions and urban morphological parameters

4.2. Application in urban wind development Once we have find out some urban morphological parameters that are highly correlated with wind potential indicators, we can apply this findings in the urban wind development. The process of the application is briefly presented in Figure 5. First of all, in the urban GIS (Geography Information System) Database of a public research centre, we have all the information of each building and each urban infrastructure element. Thus we can find out urban forms (in a neighbourhood scale) which have those urban morphological parameters with suggested high value (if it’s positive correlation with wind capacity indicator M’) or low value (if it’s negative correlation with that). With a windy urban form, we still need the local wind conditions that are above the average, to make sure that we have the right urban forms with wind potential. Then, in the angle of urban planning, the planners would give suggestions to the corresponding government department, who in consequence, if the suggestions are accepted and feasible, would give incentives to encourage multiple stakeholders to install wind turbines around their buildings and environments. The installed wind electricity would then give update data to government information centre and help adjust the incentives if somewhere is over-developed or under developed. The wind electricity produced in local community would be connected to the Smart Energy Grid, which makes the energy production and consumption efficiently and economically distributed and managed. The balance of the Smart Energy Grid will be given to the urban planning management to influence the later on energy development strategy.

22

ACCEPTED MANUSCRIPT

Fig. 5. Urban wind energy development process with finding of beneficial urban morphological parameters

5. Conclusion Built environment has a important impact on the urban wind performance. In this paper we focus on the impact of urban morphology on the urban wind potential collection. Based on a previous validation case with a reference wind tunnel experiment, the parameter settings of the CFD code are discussed and modified in order to adapt to the new model size and new computation demand. Six real urban tissues were selected for morphological analyse and wind potential evaluation. Results of different wind capacity values over different type of urban tissues and in different wind inlet directions, confirmed the great impact of urban morphology. Eleven urban morphology indicators having potential relationship with wind environment are selected for evaluation. Two wind potential indicators are taken into consideration: wind potential capacity M’ and wind potential density D'. Simulation results of the real urban tissues show that: 1) The values of M’ and D’ vary among different urban tissues. Among the six analyzed, the tissues of New York and Paris possess generally higher total wind potential capacity, and the tissues of New York and Toulouse possess generally higher wind potential density. 2) The wind performance varies from different inlet direction. Different urban tissue has its own best wind inlet direction for high wind potential collection. 3) Different wind speed thresholds upon which the exploitable wind potential is defined, shows a great impact on wind potential capacity for different urban tissue. Therefore, when selecting a wind turbine to generate electricity, the feature of cut-in speed should be an important evaluation aspect. Concerning the relationship between those morphological indicators and wind potential indicators, correlation analysis is applied. Very significant correlation relations (r > 0.88) were found between Rr, Po, FAR, λc, 𝑉𝑏 and M’, and σh/𝐻 and D'. On average, compared to the wind potential density D', the capacity M’ has relevantly higher correlation with different morphological parameters. Between the two wind thresholds conditions: U > 2m/s and U > 3m/s, with wind U > 3m/s there were higher correction coefficient on average. Therefore, with the value of some certain morphological indicators, for example, Relative rugosity Rr and the ratio between the standard deviation of the building heights and the average building height σh/𝐻, can compare quite accurately the output of wind potential over roof among different urban tissues. Without undergoing time-consuming CFD simulations of many complex urban forms, the prediction of wind potential by the morphological indicators has a great significance for the urban wind potential development. For future work, more types of urban tissues would be considered to further prove the correlation relation between the urban morphological indicators and the wind potential indicators. Renewed morphological parameters with little correlation among each other would be considered. Based on the prediction method, some highly potentiated urban tissues can be selected. Then, with local climate condition, more detailed modelling and more precise mesh generation, some careful CFD simulations can be undertaken to evaluate real wind potential in the urban environment.

Acknowledgements 23

ACCEPTED MANUSCRIPT Authors would like to thank the China Scholarship Council for funding the initial part of this research project. Thanks to North China University of Technology for offering conditions to continuing this research: Scientific Research Start-up Fund Project, School Youth Research Innovation Fund.

References

[1] United Nations. (2015). World urbanization prospects. New York: Department of Economic and Social Affairs. http://esa.un.org/unpd/wup/Publications/Files/WUP2014-Report.pdf (Accessed 21.5.2017). [2] Eurobserver'ER, Wind energy barometer 2016, Eurobserver'ER project group. (www.eurobserv-er.org) (accessed January 2017). [3] L. Aelenei, A. Ferreira, C. S. Monteiro, R. Gomes, H. Gonçalves, S. Camelo, C. Silva. Smart City: A systematic approach towards a sustainable urban transformation. Energy Procedia 91 ( 2016 ) 970 – 979. [4] F. Mosannenzadeh, A. Bisello, R. Vaccaro, V. D'Alonzo, G.W. Hunter, D. Vettorato. Smart energy city development: A story told by urban planners. Cities 64 (2017) 54–65. [5] C.F. Calvillo, A. Sánchez-Miralles , J. Villar. Energy management and planning in smart cities. Renewable and Sustainable Energy Reviews. Volume 55, March 2016, Pages 273–287. [6] N.S. Campbell, S. Stankovic, Final report of the project WEB (Wind Energy for the Built Environment), BDSP Partnership, London, 2001. [7] A.G. Dutton, J.A. Halliday, M.J. Blanch, The Feasibility of Building Mounted/Integrated Wind Turbines (BUWTs): Achieving their potential for carbon emission reductions. Final report, under contract of Carbon Trust (2002-07-028-1-6), Energy Research Reference, CCLRC, 2005. [8] WINEUR, Wind energy integration in the urban environment, Deliverable 1.1, Technology inventory report. EIE/04/130/S07.38591. Project WINEUR, 2005. [9] New Energy Huseum, Small Wind World Report 2014. Published by World Wind Energy Association, Bonn, March 2014. [10] Energy Task Force, Windmill Power for City People: A Documentation of the First Urban. ARENE Ile-deFrance, Paris, 2006. [11] Task 27 project, Small wind turbine labels development and deployment from 2008-2011, and small wind turbine in high turbulence sites from 2012-2016, International Energy Agency-Wind section. (www.ieawind.org/task_27_home_page.html) (accessed January 2017). [12] Encraft, 2009. Final report of the Warwick Wind Trials Project. (www.warwickwindtrials.org.uk) (accessed January 2017). [13] M. Moriarty, Feasibility of Small-Scale Urban Wind Energy Generation. Master's Thesis, University of Pittsburgh. 2009. [14] ADEME, Le Petit Eolien. Fiche Technique, The French Environment and Energy Management Agency. Feb 2015. [15] M. Bottema, Wind Climate and Urban Geometry. Technology University of Eindhoven, Netherlands, 1993 (Ph.D. Thesis). [16] S. Mertens, Wind energy in the built environment: Concentrator Effects of Buildings. PhD thesis, Technology University of Delft, published by Multi-Science, The Netherlands, 2006. [17] C. Beller, Urban Wind Energy. PhD thesis, Risø National Laboratory for Sustainable Energy, Technical University of Denmark, Copenhagen, 2011 (Ph.D Thesis). [18] S. Stankovic, N. Campbell, A. Harries, Urban Wind Energy. Published by Earthscan, London, 2009. [19] L. Lu, K.Y. Ip, Investigation on the feasibility and enhancement methods of wind power utilization in highrise buildings of Hong Kong. Renewable and Sustainable Energy Reviews 13(2) (2009) 450–461. [20] L. Lu, K. Sun, Wind power evaluation and utilization over a reference high-rise building in urban area. Energy and Buildings 68 (2014) 339-350. [21] K. Lynch, A Theory of Good City Form. Cambridge, MA: MIT Press, 1981. [22] A.M.K. Sharag-Eldin, Predicting natural ventilation in residential buildings in the context of urban environments. PhD thesis, University of California, Berkeley, 1998. [23] P. Massimo, C. Carola, I. B. Antoni, Climate change and urban form: Simulation studies in temperate climates. Proceedings of the International conference PLEA (Passive and Low Energy Architecture) 2015, Bologna, Italy. [24] J. Allegrini, V. Dorer, J. Carmelie, Influence of morphologies on the microclimate in urban neighborhoods. J. Wind Eng. Ind. Aerodyn. 144 (2015) 108–117. [25] C. Ratti, N. Baker, K. Steemers, Energy consumption and urban texture, Energy and Buildings 37 (2005) 762–776.

24

ACCEPTED MANUSCRIPT [26] L. Eicker, D. Monien, E. Duminil, R. Nouvel, Energy performance assessment in urban planning

competitions. Applied Energy 155 (2015) 323–333. [27] M. Grosso, Urban form and renewable energy potential. Renewable Energy I5 (1998) 331–336. [28] C. Liu, Q. Shen, An empirical analysis of the influence of urban form on household travel and energy consumption. Computers, Environment and Urban Systems 35 (2011) 347–357. [29] S. Hankey, J. D. Marshall, Impacts of urban form on future US passenger-vehicle greenhouse gas emissions. Energy Policy 38 (2010) 4880–4887. [30] S. A. Zahabia, L. Miranda-Moreno, Z. Patterson, P. Barla, C. Harding, Transportation Greenhouse Gas Emissions and its Relationship with Urban Form, Transit Accessibility and Emerging Green Technologies: A Montreal case study. EWGT 2012, Procedia - Social and Behavioral Sciences 54 (2012) 966–978. [31] E. Ng, C. Yuan, L. Chen, C. Ren, J.C.H. Fung, Improving the wind environment in high-density cities by understanding urban morphology and surface roughness: A study in Hong Kong, Landscape and Urban Planning 101 (1) (2011) 59-74. [32] N. Norte Pinto, Technologies for urban and spatial planning: virtual cities and territories. Hershey: Information Science Reference, 2014. [33] P. S. Edussuriya, Urban morphology and air quality: a study of street level air pollution in dense residential environments of Hong Kong. PhD thesis, University of Hong Kong. 2006. [34] C.H. Huang, X.N. Pham, The Influence of Community Planning on Urban Thermal Environment, 2012 International Conference on Environment Science and Engineering, IPCBEE vol.3 2(2012). [35] S. Kitous, R. Bensalem, L. Adolphe, Airflow patterns within a complex urban topography under hot and dry climate in the Algerian Sahara, Building and Environment 56 (2012) 162-175. [36] L. Sun, N. Anders, K. Jan, Effect of hilly urban morphology on dispersion in the urban boundary layer, Building and Environment 48 (2012) 195-205. [37] P. Edussuriya, A. Chan, A. Ye, Urban morphology and air quality in dense residential environments in Hong Kong. Part I: District-level analysis. Atmospheric Environment 45(27) (2011) 4789-4803. [38] R.M. Cionco, R. Ellefsen, High resolution urban morphology data for urban wind flow modeling. Atmospheric Environment Vol. 32. No. 1, pp. 7-17. 1998. [39] N. Colaninno, J.R. Cladera, K. Pfeffer, An automatic classification of urban texture: form and compactness of morphological homogeneous structures in Barcelona. ERSA Congress 2011, Barcelona, Spain. [40] J. Franke, C. Hirsch, A.G. Jensen, H.W. Krüs, M. Schatzmann, P.S. Westbury, S.D. Miles, J.A. Wisse, N.G. Wright, Recommendations on the use of CFD in wind engineering. In: van Beeck, J.P.A.J. (Ed.), Proc. Int. Conf. Urban Wind Engineering and Building Aerodynamics. COST Action C14, Impact of Wind and Storm on City Life Built Environment, von Karman Institute, Sint-Genesius-Rode, Belgium, 2004. [41] Y. Tominaga, A. Mochida, R. Yoshie, H. Kataoka, T. Nozu, M. Yoshikawa, T. Shirasawa, AIJ guidelines for practical applications of CFD to pedestrian wind environment around buildings. Journal of Wind Engineering and Industrial Aerodynamics 96 (10-11) (2008) 1749–1761. [42] B. Blocken, T. Stathopoulos, J. Carmeliet, CFD simulation of the atmospheric boundary layer: wall function problems. Atmos. Environ. 41(2) (2007) 238-252. [43] B. Blocken, W.D. Janssen, T. van Hooff, CFD simulation for pedestrian wind comfort and wind safety in urban areas: General decision framework and case study for the Eindhoven University campus. Environmental Modelling & Software 30 (2012) 15-34. [44] Y. Yang, Relationship between urban morphology and wind conditions in ideal wind model. The 18th International Seminar on Urban Form, Montréal. 2011. [45] L. Lu, K.Y. Ip, Investigation on the feasibility and enhancement methods of wind power utilization in highrise buildings of Hong Kong. Renewable and Sustainable Energy Reviews 13 (2) (2009) 450–461. [46] D.F. Zhang, Wind Energy Effective Utilization in the Built Environment, Shandong Jianzhu University, Ji Nan, 2010 (MSc Thesis, Chinese). [47] A. Baskaran, A. Kashef, Investigation of air flow around buildings using computational fluid dynamics techniques. Engineering Structures, 18 (11) (1996) 861–873, 875. [48] L. Lu, K. Sun, Wind power evaluation and utilization over a reference high-rise building in urban area. Energy and Buildings 68 (2014) 339–350. [49] B. Blocken, J. Carmeliet, T. Stathopoulos, CFD evaluation of the wind speed conditions in passages between buildings – effect of wall-function roughness modifications on the atmospheric boundary layer flow. Journal of Wind Engineering and Industrial Aerodynamics, 95 (9–11) (2007) 941-962. [50] Y. Zhang, Wind-Energy Efficiency Study and Structural Analysis of Building Integrated / Mounted Wind Turbines, Zhejiang University, Hangzhou, 2011 (Msc Thesis, Chinese).

25

ACCEPTED MANUSCRIPT [51] R. Yang, Wind Energy Utilization on the Roof of the High-rise Building, Huaqiao University, Quanzhou, 2011 (MSc Thesis, Chinese). [52] L. Campos-Arriaga, Wind Energy in the Built Environment: A Design Analysis Using CFD and Wind Tunnel Modelling Approach, Published PhD thesis, University of Nottingham. 2009. [53] X. Shi, Y.Y. Zhu, J.Duan, R.Q. Shao, J.G.Wang, Assessment of pedestrian wind environment in urban planning design. Landscape and Urban Planning 140 (2015) 17–28. [54] S. Srifuengfung, W. Peerapun, Investigation of the ventilation rate around different urban morphological property types: High rise-VS-Low rise in Bangkok’s high density areas. ABAC Journal 33 (3) (2013) 65-81. [55] I. Janajreh, L. Su, F. Alan. Wind energy assessment: Masdar City case study. Renewable Energy 52 (2013) 815. [56] J.T. Millward-Hopkins, A.S. Tomlin, L. Ma, D.B. Ingham, M. Pourkashanian, Assessing the potential of urban wind energy in a major UK city using an analytical model. Renewable Energy 60 (2013) 701-710. [57] B. Wang, Les Impacts de la Morphologie Urbaine sur le Vent: Performance énergétique éolienne à l'échelle du quartier. Thèse de Université de Toulouse. Avril 2015. [58] R. Yoshie, A. Mochida, Y. Tominaga, H. Kataoka, K. Harimoto, T. Nozu, T. Shira-sawa, Cooperative project for CFD prediction of pedestrian wind environmentin the Architectural Institute of Japan, J. Wind Eng. Ind. Aerodyn. 95 (2007) 1551–1578. [59] B. Wang, L.D. Cot, L. Adolphe, S. Geoffroy, J. Morchain, Estimation of wind energy over roof of two perpendicular buildings. Energy and Buildings 88 (2015) 57–67. [60] H. Yoshida, M. Omae, An approach for analysis of urban morphology: methods to derive morphological properties of city blocks by using an urban landscape model and their interpretations. Computers, Environment and Urban Systems 29 (2005) 223–247. [61] Grimmond, C. S. B. and Oke, T. R., Aerodynamic Properties of Urban Areas Derived from Analysis of Surface For m, J. Appl. Meteorol. 38 (1999) 1262–1292. [62] L. Adolphe, A simplified model of urban morphology: application to an analysis of the environmental performance of cities. Environment and Planning B: Planning and Design, Volume 28, (2001)183-200. [63] C. Balocco, G. Grazzini, Thermodynamic parameters for energy sustainability of urban area. Solar Energy 69 (4) (2000) 351–356. [64] L. Adolphe, (participate with the TMU, the ABC group, l’IRPHE, and the Laboratoire de Psychologie de l’Environnement –Paris V). SAGACités: Vers un Système d’Aide à la Gestion des Ambiances Urbaines. Rapport final, SAGACités. Ecoles d’Architecture de Toulouse et de Bordeaux. Fév. 2002. [65] X. Rochel, La Morphologie Urbaine, Université de Nancy II, Foxit Software Company, 2007. [66] Google company. Google maps. (www.googlemaps.com) (accessed January 2017).

26

ACCEPTED MANUSCRIPT Highlights 

Potential windy urban forms were evaluated and selected.



Potential urban morphological indicators having interaction with wind environment were selected.



Wind energy potential over roof of six real urban tissues was evaluated and compared.



Several urban morphological indicators were found with high correlation with wind energy indicators.

.