Available online at www.sciencedirect.

com

ScienceDirect
Available online at www.sciencedirect.com
Energy Procedia 00 (2017) 000–000
www.elsevier.com/locate/procedia
ScienceDirect
Energy Procedia 134 (2017) 757–767

9th International Conference on Sustainability in Energy and Buildings, SEB-17, 5-7 July 2017,
Chania, Crete, Greece

An Optimization design Approach of Football Stadium Canopy

Forms Based On Field Wind Environment Simulation
Ligang Shi a, *, Rongrong An a,b
a
Harbin institute of technology, School of Architecture,Heilongjiang Cold Region Architectural Science Key Laboratory , Harbin,150080,China
b
Sunac China Holdings Limited,Tianjin,300384, China

Abstract

After the issue of General Plan of Chinese Football Reform in 2015, the development and construction of football stadium in
China was put on an important agenda. Field wind environment will make a serious impact on players’ performance and game
result，while canopy form plays great effect on field wind environment. The primary intention of this research is to reveal the
relationship between field wind environment and canopy forms based on CFD simulation of wind environment, then construct an
optimization design approach of football stadium canopy form . STAR-CCM + 9.04.009 is set as simulation platform in this
paper. As for the current developing status of football game in China, a 30,000-capacity medium-sized football stadium model is
proposed for research object. And then CFD simulation is carried out on the canopy section profiles and the canopy connection
permeability, different external wind speed such as 5 m/s, 10 m/s, 15m/s ,flows into the model from east side in all the cases.
Furthermore, the wind field distribution uniformity and wind speed variation stability are taken as the evaluation criterion. In this
way,the impact mechanism which different canopy forms play on the field wind environment in football stadium is attempted to
characterize from the experimental results. Finally the proposal of the football stadium optimal design strategies are put forward .
© 2017 The Authors. Published by Elsevier Ltd.
© 2017 The under
Peer-review Authors. Published by
responsibility of Elsevier Ltd.
KES International.
Peer-review under responsibility of KES International.
Keywords: football stadium; canopy forms; field wind environment; CFD simulation

1. Introduction

After the issue of General Plan of Chinese Football Reform in 2015, the development and construction of football

* Corresponding author. Tel.: +86-0451-86281142; fax: +86-0451-86281142.

E-mail address: slg0312@hit.edu.cn

1876-6102 © 2017 The Authors. Published by Elsevier Ltd.

Peer-review under responsibility of KES International.

1876-6102 © 2017 The Authors. Published by Elsevier Ltd.

Peer-review under responsibility of KES International.
10.1016/j.egypro.2017.09.575
758 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767
Ligang Shi,Rongrong An/ Energy Procedia 00 (2017) 000–000

Table 1. Some cases that strong wind affected the results of football matches in recent years
Time Place Both sides Beaufort Wind Scale Results
2006-03-13 Cagliari,Italy Florence VS Cagliari 10 Players injured, game interrupted
2012-01-22 Alkmaar,Netherlands Alkmaar VS Ajax 6 [2]
Alkmaar lost top spot
2012-03-01 Haifa, Israel Maccabi Haifa VS Dynamo Kyiv About 7 [3]
Resulting own goal
2012-10-28 Marseille, France Marseille VS Lyon About 7 [4]
match postponed
2013-11-3 Stoke,UK Stoke City VS Southampton 6 Most long-distance goals in
[2]
Premier League history
2014-12-10 Basel, Switzerland Basel VS Liverpool 6 Liverpool was eliminated in the
[2]
champions league
2014-02-13 Manchester,UK Manchester city VS Sunderland 7 [5]
match Postponed
2014-02-13 Liverpool,UK Everton VS Crystal Palace 6 [5]
match Postponed
2017-02-8 Vigo,Spain Celta de Vigo VS Real Madrid 7 [6]
match postponed
2017-02-8 La Coruna,Spain Deportivo La Coruna VS Betis 7 [7]
match postponed

stadium in China was put on an important agenda.[1] As an outdoor sport, football is affected by the weather
inevitably. Because of the relationship between resistance and quality, wind plays a great influence on the football
trajectory and velocity. The field wind environment plays an important impact on players’ performance and game
results(Tab. 1). FIFA stipulates the standard football weight is 410-450 grams at the beginning of the game.[8] Under
the influence of the wind, the football speed and flight trajectory is easy to change. In addition, the wind will cause
obstruction to the referee's sight line, thus affecting their right judge the facts related to the game. Canopy forms is
an important factor affecting the field wind environment. In this paper, we will discuss the influence of different
canopies on the wind environment in football field, which provide design reference for football stadium construction.

2. Wind influence on football stadium design

A design process generates a description of a design object which satisfies a given set of design requirements and
fulfils a given set of design process objectives. During the design process, both the partial description of the design
artefact and the design requirements and process objectives change, often evolving from an abstract definition
toward measurable criteria. [9] In earlier phases of design, assessing the fulfilment of design requirements relies on
the insight of the designer and focuses on a limited range of performances (like functional and esthetics). For other
performances (like most of the engineering related aspects), the assessment is usually postponed. Considering the
impact choices made during conceptual design have on the success of the design solution. Enlarging the set of
performances assessed at an early stage, enhances inter-disciplinary ,and creates a visual link between form and
numeric performance evaluations, which can reduce the investment in poor performing solutions. [10]Wind
performance based building design has received quite some attention in the past, especially in the arenas. A
computational fluid dynamics (CFD) mathematical method is usually applied for wind flow analysis.[11-14]Three-
dimensional steady state RANS(Reynolds-averaged Navier-Stokes) CFD simulations were used in combination
with the new Dutch wind nuisance standard to assess pedestrian wind comfort around a large football stadium in
Amsterdam, before and after the addition of new high-rise buildings. [15] RANS model simulations was also
presented to determine the wind flow pattern to evaluate the influence of both overall stadium geometry and roof
slope on the area of the stand that is wetted by wind-driven rain, which provides some design guidelines to avoid
this type of spectator discomfort. [16-17] Isothermal CFD simulations of coupled urban wind flow and indoor natural
ventilation were presented to assess the influence of wind direction and urban surroundings on the air change rate
per hour(ACH) of a large semi-enclosed stadium. [18] CFD mathematical models were developed using ANSYS
Fluent software to evaluate the wind flow patterns around and inside the future football stadium retrofit design in
West Lafayette, IN. [19] CFD simulations for a prototype of a railway station shelter were conducted to predict the
local wind velocities in and around the shelter. [20] Ventilation analysis was conducted by CFD simulations towards
the development of optimised conditioning strategies. [21] Several CFD simulations of the three-dimensional airflow
with heat and mass transfer in a complex geometry indoor swimming pool were performed to evaluate the weather
effect on the indoor atmosphere. [22] Recent validation efforts of CFD for wind performance studies for these
 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767 759
Ligang Shi,Rongrong An / Energy Procedia 00 (2017) 000–000

applications have provided confidence to extend CFD simulation to other applications, which assists architects and
engineers in finding solutions for a safe and efficient design for a stadium. A quick and straightforward assessment
is often required in football stadium canopy design, especially from the perspective of field wind environment,
which is a weak link of the environment evaluation system and crucial to the football match. To meet these
requirements, CFD simulations are performed to investigate the impact of roof geometry on field wind environment
in football stadium, football stadium optimal design strategies are presented based on wind simulation in this paper.

3. Methodology

3.1. 3d simulation model

As for the current developing average attendance status of football game in China(Fig.1), a 30,000-capacity
medium-sized football stadium is proposed for research object, which is the most popular square round building
shape. The model factors are canopy section profile and canopy connection permeability.

Fig. 1. Total and Average attendance Recent Trend in China Super League.

Through the world-wide investigation of 80 famous football stadia, three main categories is classified from the
perspective of canopy plan form, that is four-side-connection type, four-side-independent type and double-sides type,
among them, four-side-connection type and four-sides-independent type has an absolute advantage of 87%, double-
side type only accounts for 9%,the rest holds 4%. [2] The research content of this paper is focused on the first two
types canopies.
The canopy section profile can be divided into three categories ,that is upward, flat and downward. which plan is
the same，four-independent canopy model size is 196m × 159m, epitaxial chamfer radius is 36m. The size of long
canopy is 121m × 34m,short canopy is 84m × 34m. From the section, the highest point of the stand is 23m,the
thickness of canopy is 1m. The vertical height between the canopy and the top of the stand is 5m. The upward and
downward canopy are inclined up and down by 5° respectively on the basis of flat canopy.
The canopy connection permeability model can be divided into fully-enclosed connection, semi-closed
connection , full-open connection three types,which plan is the same，four- through canopy model size is the same
as the three models in canopy section profile. There is no gap between the seat and the canopy in the fully-enclosed
connection .The semi-closed connection appears a porosity of 50%.That is, a 10 m-long partition is provided at
intervals of 10 m.The full-open connection keeps the gap between the seat and the canopy.
760 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767
Ligang Shi,Rongrong An/ Energy Procedia 00 (2017) 000–000

Table 2. Models extraction

Plan Section Perspective

upward canopy

Canopy
flat canopy
section profile

downward canopy

fully-
enclosed connection

canopy connection semi-closed

permeability connection

full-open connection

3.2. Models in STAR-CCM+

According to open-reviewed results of several commonly used CFD simulation softwares of Japan Society of
Automotive Engineers (JSAE) , vortex separation and shedding which STAR-CCM+ calculated is most accurate,
the simulation results are very close to the experimental data. Compared with the traditional tetrahedral mesh,
polyhedral mesh which STAR-CCM+ promotes has the advantages of high precision and fast partition. [23] Therefore
STAR-CCM + 9.04.009 is set as simulation platform in this paper. The calculation domain is set as 700m × 500m ×
60m in this paper. The basin setting satisfies the requirement that the blocking rate is less than 3%.The domain body
is divided into polyhedral meshes. A five-layer prism grid with a thickness of 20% is used at the boundary
intersection. The mesh needs to be encrypted near the football stadium area. The calculation domain is divided
unstructured grid, generating 450-550 thousands units which depends on different models(Fig.2).
The case is a model of flow around the single object. The Realizable k-ε turbulence model is adopted in
thispaper.[8] The inlet boundary is the velocity，while the outlet boundary is the pressure. The rest boundaries is
symmetrical plan. The floor and the building are set as slip walls. Selects 3 initial wind speeds，which are 5 m/s, 10
m/s and 15 m/s at height of 10m，adopting the urban gradient wind parameter. The expression of gradient wind is
n
u z
= 
u1  z 1 
 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767 761
Ligang Shi,Rongrong An / Energy Procedia 00 (2017) 000–000

Z1 refers to height 10m，and U1 is velocity at reference height 5m/s，10m/s，15m/s. Z is a certain height，and U

is the velocity corresponding to this height. N represents the ground roughness, which is set as 0.15 according to
China's "building structural load specifications" (GB50009-2012).[24]The wind outside flows into the football
stadium vertical to the long axis.

Fig. 2. Schematic diagram of the grid.

4. Experimental results and analysis

FIFA does not specify the maximum wind speed for football matches. But sometimes windy weather interrupts
the football game. The purpose of this paper is to explore suitable canopy forms that can create the best field wind
environment.In this paper, the evaluation principle is the wind field distribution uniformity and wind velocity
stability.In addition, there is an auxiliary analysis principle "reducing wind speed"for windy areas particularly.[2]

4.1. Experimental results and analysis of the canopy profile model

（1）Experimental results and analysis of horizontal section

In this paper, the external winds flow into the model from east side in all the cases. The direction of all the
following horizontal sections is the same as Fig.3.That is, the north side is right. To avoid redundancy, the article is
no longer place compass one by one .

Fig. 3. Wind direction schematic diagram.

Upward canopy：When the inlet wind velocity is 5m/s, two vortices complicate the whole wind field.The north-
south wind field is divided into three distinct regions.The wind velocity on both sides of field is lower while the
middle is high.When the inlet wind velocity is 10 m/s，there are significant vortices in both east side of the goal
area. The highest wind speed is about 6.3m/s ,which appears in the west of the center, The lowest point is close to
0m / s,which occurs on both sides of the goal area. When the inlet wind velocity is 15m/s, a local vortex appears at
the northeast corner.The wind speed of part of the area is below 5.7m/s,including the corner area and the goal area.
And the wind field distributes relatively uniform in other regions.The wind speed in most regions is between 8.5m/s
and 14.1m/s.
762 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767
Ligang Shi,Rongrong An/ Energy Procedia 00 (2017) 000–000

Table 3. Horizontal section experimental results of canopy profile model

5m/s 10m/s 15m/s

upward
canopy

flat canopy

downward
canopy

Flat canopy：When the inlet wind velocity is 5m/s,the wind speed reduces gradiently from west to east.The
maximum wind speed is 2.8m/s, the lowest is 0.3m/s.When the inlet wind velocity is 10m/s，the wind speed
changes slowly in the main body area.Wind speed is between 5.0 m/s and 6.6 m/s.The wind speed gradually
decreases to 0.55m/s in the southeast area.The wind speed reduces sharply to zero in the northeastern side of the
region.When the inlet wind velocity is 15m/s, Wind speed is slow in the southeast region. And the wind field
distribution is relatively uniform in other regions , wind velocity varies mainly from the 4.9m / s to 9.9m / s . The
highest wind speed is 11.9m / s, which is in the southwest area.
Downward canopy: When the inlet wind velocity is 5m/s, violent vortexs appear in the north and south ends of
the middle line.Although the full-field wind speed difference is only about 2.5m/s，wind speed changing rapidly in
the main area ,which has a bad effect on football game.When the inlet wind velocity is 10m/s, a high wind speed
region appears near the center line and the east side line，varying between 5.1 m/s and 6.9 m/s. Wind speed is lower
than 3.4 m/s in other regions.When the inlet wind velocity is 15 m/s, small-scale vortex appears in the north goal
region.The wind speed in two goal areas and the west line area is relatively low，which is below 2.7m/s.The
difference of the full-field wind speed is only about 6.8m/s.Full field wind speed transites relatively flat.

（2）Experimental results and analysis of average wind speed

The left histogram indicates the 1.5m cross-section average wind speed when the inlet wind speed is 5m/s, 10m/s
and 15m/s respectively. It can be shown from the right-hand trend diagram that the inclination angle of the
downward canopy to the x-axis is the smallest.And there is a tendency that the angle is smaller as the wind speed
increases.
 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767 763
Ligang Shi,Rongrong An / Energy Procedia 00 (2017) 000–000

Fig. 4. Average wind speed comparison of the canopy profile models.

4.2. Experimental results and analysis of the canopy connection permeability

（1）Experimental results and analysis of horizontal section

Table 4. Horizontal section experimental results of canopy connection permeability

5m/s 10m/s 15m/s

Fully- enclosed
connection

Semi- closed
connection

Full-open
connection

Fully- enclosed connection：When the inlet wind velocity is 5m/s, wind speed is about 3.3 m/s in most western
region and decreases gradually from west to east, which drops to under 0.8m/s in the east sideline.The wind speed of
this model is high and change rapidly.There are intense vortexes in southeast and northeast corner， causing
obstacles to passing and corner kicks.When the inlet wind velocity is 10m/s, the wind speed is slightly lower near
the eastern edge，between 1.2 m/s and 4.8m/s.The wind speed in more than 50% pitch region is above 7.0 m/s.It
shows that the fully-enclosed connection has little effect on reducing wind speed.When the inlet wind velocity is
764 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767
Ligang Shi,Rongrong An/ Energy Procedia 00 (2017) 000–000

15m/s, full field wind speed is too high.The wind speed is more than 9.5 m/s in 80% -90% area.Wind speed changes
quickly and without transition.If we use this model in windy area, football game will face a serious threat.
Semi-closed connection：When the inlet wind velocity is 5 m/s, wind speed in west area and the central of east
area is relatively high，about 1.7-2.6 m/s .The wind speed decreases gradually in the southeast and northeastern
regions，dropping to 0.3 m/s.Full field wind speed transits relatively flat，and the difference between the highest
and lowest points is about 2.3 m/s.When the inlet wind velocity is 10m/s, obvious vortexes appear in the southeast
and northeastern region.In general, it is divided into three high wind speed zones (southwest, northwest, east central)
and two low wind speed zones (southeast and northeast).When the inlet wind velocity is 15m/s, there are two high
wind speed zones in the southwestern and northwest regions.The maximum wind speed is about 8.4 m/s and 7.2
m/s .The minimum of wind speed is about 0.8m/s, appearing in the south goal region.In this model,wind speed is
mainly a gradient change，but the relatively complex wind field will have a negative impact on the football game.
Full-open connection：When the inlet wind velocity is 5 m/s, there is no abrupt change in wind speed.Wind
flows smoothly and no vortex appears.The wind speed peaks at about 2.3 m/s and troughs close to 0.8 m/s, the
difference is only 1.5 m/s,which is suitable for football match.When the inlet wind velocity is 10 m/s, two high wind
speed regions appear in the northwest and southwest of the pitch，about 4.5 m/s. Wind speed is lower in the
southeast and northeast region. The maximum wind speed difference is only about 3.8 m/s. When the inlet wind
velocity is 15 m/s,field wind speed decreases gradually from west to east，reaching trough to 0.6 m/s in the
southeast corner .A low-velocity region appears in the west-central region，which varies from 4.2 m/s to 4.8
m/s.Generally speaking,full-open connection model can generate a relatively uniform and stable wind field,which is
more suitable for football match.

（2）Experimental results and analysis of average wind speed

Under three different inlet wind speeds , the average field wind speed of the fully-enclosed connection is much
higher than the other two models.It can be seen from the trend graph that the field wind speed in fully-enclosed
connection model changes fastest ,and it changes slowest in the semi-closed connection model when inlet wind
speed increases. From the perspective of "wind velocity stability" ,semi-closed connection model performs best and
the fully-enclosed connection model performs the worst.

Fig. 5. Average wind speed comparison of the canopy connection permeability models.

5. Discussion and design recommendations

5.1. Design recommendations for canopy section profile

Wind field distribution uniformity: As the angle between the downward canopy and the seat is smaller, there is
strong turbulence,which leads to complex wind field.In the comparison of field wind environment distribution
uniformity, the upward canopy performs better than the downward canopy.
 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767 765
Ligang Shi,Rongrong An / Energy Procedia 00 (2017) 000–000

Table 5. The brief results for canopy section profile

Distribution uniformity Maintaining stability Reducing wind speed(Windy areas)

upward canopy Better Poor Poor
flat canopy Moderate Moderate Moderate
downward canopy Poor Better Better
Wind velocity stability：From the perspective of maintaining stability of the field wind environment，the
performance of the downward canopy is better than the upward canopy.
Reducing wind speed (Windy areas)：At the same inlet wind speed, the average wind speed of the upward
canopy is higher than that of the downward canopy.
When constructing football stadium in the windy area，if other conditions have little effect，it is recommended
to use the downward canopy and avoid using upward canopy.Many top european football clubs’ home fields apply
the downward canopies such as London Emirates Stadium(Fig.6)，Frankfurt Commerzbank-Stadium etc.

Fig. 6.Profile of Emirates Stadium ，London.

5.2. Design recommendations for canopy connection permeability

Table 6. The brief results for canopy connection permeability

Distribution uniformity Maintaining Stability Reducing wind speed(Windy areas)

Fully- enclosed connection Poor Poor Poor
Semi- closed connection Moderate Better Moderate
Full-open connection Better Moderate Moderate
Wind field distribution uniformity: The full-open connection model has the smallest difference of full-field wind
speed，performancing the best.The wind speed of semi-closed connection model changes step by step,which is
better than that of the fully-enclosed connection model ,the latter is affected by vortices，performancing the worst.
Wind velocity stability：From the perspective of maintaining stability of the field wind environment，the best is
the semi-closed connection model, the worst is fully- enclosed connection model.
Reducing wind speed (Windy areas)：From reducing the inlet wind speed and maintain low field wind speed，
semi-closed and full-open connection model is similar.The fully-enclosed connection model is far inferior to the
other two models.
When designing football stadium in windy area，if other conditions have little effect，we suggest to avoid using
the fully-enclosed connections model.While the semi-closed and fully-open connection model can both maintain
appropriate internal wind environment.Closure shouldn’t be provided at the gap between the canopy and the top seat,
only retain the support structure,or open the window here，trying to avoid forming a completely closed connection.
Many well-known football stadium use full-open connection between the canopy and seat.Such as Seoul World Cup
Stadium in Korea and Commerzbank-Arena in Germany（Fig.6）.
766 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767
Ligang Shi,Rongrong An/ Energy Procedia 00 (2017) 000–000

Fig. 6. Commerzbank-Arena，Frankfurt.

6. Conclusions

As an effective method ,CFD simulation is used to investigate the relation between the canopy form and field
wind environment. A 30,000-capacity football stadium model is proposed for experiment. Based on the field wind
environment simulation performance, comprehensively speaking, it is recommended that the priority order of
medium-sized football stadium canopy section profile is as following order: firstly the downward canopy,then flat
canopy,and thirdly the upward canopy.As for the canopy connection permeability, from the performance of maintain
appropriate internal wind environment, the semi-closed and fully-open connection model is better than the fully-
enclosed connections model.

Acknowledgements

Project Supported By National Natural Science Foundation of China (Grant No.51678180)

References

[1] General Plan of Chinese Football Reform. Beijing：People’s publishing house. 2015: p.15-21.
[2] An Rongrong. research on football stadium canopy shape based on the field wind environment simulation. Harbin Institute of
technology.2016.
[3]Jimmy Nsubuga. Maccabi Haifa keeper scores bizarre wind-assisted own goal. ( Mar 1,2012)http://metro.co.uk/2012/03/01/maccabi-haifa-
goalkeeper-assaf-mendes-gets-wind-assistance-to-score-bizarre-own-goal-337040/
[4] Daniel Quinton-Navarro. The highly anticipated top-of-the-table contest has been cancelled by the Departmental Security Committee because
of the windy conditions in the region on Sunday.( Oct 29, 2012) http://www.goal.com/en-sg/news/3868/main/2012/10/29/3484804/marseille-
lyon-clash-postponed-due-to-bad-weather
[5] Matches at the Etihad Stadium and Goodison Park were both called off due to the severe storm sweeping across Britain, halting one fan's 30-
year wait to see Everton play live (Feb 13, 2014) http://www.goal.com/en-sg/news/3991/premier- league/2014/02/13/4615418/weather-
wreaks-havoc-on-premier-league-fixtures
[6]High winds and rain batter Spain and force cancellation of Celta de Vigo vs Real Madridmatch .(Feb 4, 2017)
http://www.thinkspain.com/news-spain/28597/high-winds-and-rain-batter-spain-and-force-cancellation-of-celta-de-vigo-vs-real-madrid-
match
[7]Associated Press . deportivo-betis-match-postponed-by-wind-damage.( Feb3,2017) http://www.washingtontimes.com/news/2017/feb/3/
deportivo-betis-match-postponed-by-wind-damage-to-/
[8] FIFA. Laws of the game.Switzerland.2016: p.10-15.
[9] P.H.G. van Langen, F.M.T. Brazier, Design space exploration revisited, in: Artificial Intelligence for Engineering Design, Analysis and
Manufacturing, vol. 20 (Issue 2, Special Issue: Design Spaces: The Explicit Representation of Spaces of Alternatives), 2006.
[10] Michela Turrin , Peter von Buelow, Rudi Stouffs. Design explorations of performance driven geometry in architectural design using
parametric modelling and genetic algorithms. Advanced Engineering Informatics.2011 (25), p.656-675
[11] Reiter, S. Assessing wind comfort in urban planning. Environment and Planning B: Planning and Design 2010, 37 (5), p.857-873.
[12] Eric N Jal. USING CFD FOR SPORTS ARENA AND STADIA DESIGN. EcoLibrium.2004(10):20-25
[13] Bert Blocken.50 years of Computational Wind Engineering: Past, present and future. Journal of Wind Engineering and Industrial
Aerodynamics. 2014(129): 69-102
[14] M.Turrin, D.Yang, A.D’Aquilio, R.Sileryte, Y.Sun.Computational Design for Sport Buildings. Procedia Engineering .2016(147 ) 878 – 883
[15] B. Blocken, J. Persoon. Pedestrian wind comfort around a large football stadium in an urban environment: CFD simulation, validation and
application of the new Dutch wind nuisance standard. Journal of Wind Engineering and Industrial Aerodynamics. 2009, 97 (5–6), p.255-270.
[16] T. van Hooff, B. Blocken, M. van Harten. 3D CFD simulations of wind flow and wind-driven rain shelter in sports stadia: Influence of
stadium geometry. Building and Environment. 2011 (46). p. 22-37.
[17] J. Persoon, T. van Hooff, B. Blocken, J. Carmeliet, M.H. de Wit. On the impact of roof geometry on rain shelter in football stadia. Journal
of Wind Engineering. and Industrial Aerodynamics. 2008 (96) . p.1274-1293
 Ligang Shi et al. / Energy Procedia 134 (2017) 757–767 767
Ligang Shi,Rongrong An / Energy Procedia 00 (2017) 000–000

[18]T. van Hooff, B. Blocken. On the effect of wind direction and urban surroundings on natural ventilation of a large semi-enclosed stadium.
Computers & Fluids. 2010 (39). p. 1146-1155.
[19] Weijun Mei , Ming Qu. Evaluation and Analysis of Wind Flow for a Football Stadium. Procedia Engineering. 2016 , 145. p.774-781.
[20] Jonas Allegrini, Aytaç Kubilay.Wind sheltering effect of a small railway station shelter and its impact on wind comfort for passengers.
Journal of Wind Engineering & Industrial Aerodynamics.2017(164): 82–95
[21] Priyadarsini Rajagopalan, Mark B. Luther. Thermal and ventilation performance of a naturally ventilated sports hall within an aquatic
centre.Energy and Buildings.2013(58):111-122
[22] Abdelhakim Limane, Hachimi Fellouah, Nicolas Galanis. Simulation of airflow with heat and mass transfer in an indoor swimming pool by
OpenFOAM. International Journal of Heat and Mass Transfer. 2017(109): 862-878
[23] STAR-CCM+|CD-adapco[EB/OL].（2015-5-9）[2016-12-10].http://www.cd-adapco.com/ zh-hans/products/star-ccm.
[24] Ministry of housing and urban-rural development of People’s Republic of China. Load code for the design of building structures(GB 50009-
2012). Beijing ：China Architecture & Building Press.2012.