- Email: [email protected]
Generic models for pedestrian-level winds in built-up regions
JOURNAL OF
ELSEVIER
Journal of Wind Engineering and Industrial Aerodynamics54/55 (1995) 515-525
~/~1/~
Generic models for pedestrian-level winds in built-up regions Theodore Stathopoulos*, Hanqing Wu Centre for Building Studies, Concordia University, 1455 de Maisonneuve Blvd. W., Montreal Quebec, Canada H3G 1M8
Abstract
Wind conditions over streets in built-up cities were examined in a boundary-layer windtunnel model study. Presented in this paper are results of wind speeds affected by a number of parameters such as the spatial density of street blocks, the building height over surroundings, the relative location of buildings and the direction of approaching wind. Based on literature information and current findings, some generic models have been established with the empirical relations between wind conditions and building configurations. These models can be applied for the preliminary estimation of pedestrian-level wind environmental conditions in built-up regions.
1. Introduction
Wind conditions over streets are an increasing concern for pedestrian comfort, pollutant dispersion, snow accumulation, etc. Such conditions depend on meteorological data, upstream terrains, surrounding blocks and adjacent buildings under consideration. Wind-tunnel experiments have been carried out for simplified models such as isolated buildings, twin buildings and more complex building combinations [1, 2-1. For several simple models, parametric relations have been attempted between wind speeds and building configurations including dimensions, geometries and relative positions. However, those relations cannot directly apply for practical situations unless the influence of surroundings and other impact parameters is properly taken into consideration. Extensive wind-tunnel experiments were performed by Murakami et al. [3] for tall buildings surrounded by uniformly distributed square blocks. The impact on wind flows was investigated with respect to building height, street width, surrounding height, wind direction and other parameters. A similar experiment was reported by
*Corresponding author. 0167-6105/95/$09.50 © 1995 ElsevierScienceB.V. All rights reserved SSD1 0 1 6 7 - 6 1 0 5 ( 9 4 ) 0 0 0 6 8 - O
516
T. Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
Jamieson et al. [4], concentrating on the effect of architectural details of the tall buildings. The Wiren study [5] was carried out for a group of nine buildings with average and maximum wind speeds presented as functions of wind directions. In studies reported by Isyumov et al. [6] and Livesey [7], street blocks were constituted of several buildings of different heights so that the distribution and the orientation of large buildings within the block became influential to wind patterns over adjacent streets. In order to generalize the knowledge of wind conditions over streets in built-up cities, research has been carried out on (1) the measurement of wind speeds along streets between uniform blocks; (2) the evaluation of wind effects of tall buildings surrounded by uniform street blocks; and (3) the analysis and synthesis of the results obtained from generic models which simulate actual city configurations. This paper presents details of the wind-tunnel experiments and the generic models for wind conditions over streets. Although the application of the established models may be limited at the present stage, these generic models and associated empirical relations are a useful progress towards some preliminary assessment of wind conditions in the urban environment.
2. Experimental setup Experiments were conducted in a boundary-layer wind tunnel designed for building aerodynamics applications. Suburban wind conditions were simulated with a mean speed profile of power-law exponent equal to 0.25 and a geometric scale of 1:500. Building models were made of wood pieces with the same dimensions of 19x 1 0 0 x 2 0 0 m m , corresponding to full-scale street blocks 9.5m high and 50 x 100 m in plan size. The block height could be increased easily by piling up wood pieces, as shown in Fig. 1. In a typical model arrangement, there were usually five rows of blocks upstream of the main building which was located at the centre of the turntable of wind tunnel. In the current study, both the block height h and the main building height H were multiples of 9.5 m, the plan dimensions W and D were either 100 m or 50 m depending on the wind direction, the width of along-wind streets (LA) was 25 m or 15 m whereas that of cross-wind streets (Lc) was kept constant at 25 m. Irwin's surface wind sensors [8] were utilized considering their advantages in measuring a large number of horizontal wind speeds - see also the additional work carried out on these sensors [9]. With the height of 4 mm (2 m in full scale) above the ground, 37 sensors were distributed as shown in Fig. 2 for street widths of 25 m. The least distance between sensors was 20 mm in order to avoid any possible interference. The sensors were installed in the centre of the streets and at distances of 5 mm (2.5 m in full scale) from buildings, mainly on sidewalks and crosswalks of streets where both high-speed winds and pedestrian activities usually take place. While most attention was paid to the along-wind street, the wind speeds on cross streets were also measured on crosswalks, as well as in front of and behind the main building. Such a sensor arrangement also applied for the street models with the short side (50 m) facing the oncoming wind. When a narrow street (LA = 15 m) was tested, six measurement points would be covered by building models.
T Stathopoulos, H. Wu/J. WindEng. Ind. Aerodyn. 54/55 (1995) 515-525
517
D
Wind
LA
~~L c D
Fig. 1. A typicalmodel arrangementwith a tall buildingsurrounded by uniformstreet blocks.
Lc
ooI
D
•
W
Fig. 2. Distribution of measurementpoints on streets 25 m wide.
In this study, mean wind speeds (U) are normalized by two reference speeds, namely, U0, the mean wind speed at each measured point when no building models are present, and Us, the wind speed on streets between uniform blocks when all street blocks are of the same height (h = H). By definition, U/Us = 1 at all measured points when H/h = 1. For the same building configuration, values of U/Uo are usually lower than one due to the sheltering effect caused by buildings.
3. Blockage ratio
For uniform street blocks, wind speeds along streets are mainly dependent upon the distribution and density of street blocks. If buildings on a street block are typified by a single cuboid as shown in Fig. 3, the following two area ratios can be defined according to Oke [10]: (1) the plan density, Art/AL, which is the ratio of the roof area
518
T. Stathopoulos, H. Wu/J. WindEng. Ind. Aerodyn. 54/55 (1995) 515-525
AF i
D
:
Fig. 3. A sketch of simplifiedstreet blocks: dimensionsand areas.
of building over that of the building lot; and (2) the roughness density, As/AL, based on the front area of the building. These ratios can be expressed as follows,
AR
WD
As
Wh
AL
(W + LA) (D + Lc)'
AL
(W + LA) (D + Lc)'
(1)
where W and D are cross- and along-wind lengths of a block, Lc and LA are widths of cross- and along-wind streets around the block, and h is the block height. The two ratios have the same value if the street block is a cube, i.e. W = D = h. The roughness density is related not only to the plan dimensions, but also to the building height; it may be more indicative than the plan density from the viewpoint of wind conditions on streets. However, both ratios are based on the lot area, that, sometimes, may not be quite suitable for the representation of street densities and wind intensities. For non-cubic blocks, an increase in the along-wind length (D) will, by definition, reduce the roughness density, but may increase the plan density in certain cases. On the other hand, it is well known that the effect of along-wind length is insignificant on the wind speeds around building corners and between two side-by-side buildings. It is windward building surfaces that affect the oncoming wind flow and influence the wind conditions at street level. Based on these considerations, a new area ratio using only windward areas of buildings and streets is proposed here, As
R~ - AF
Wh
(W + LA) 2 '
(2)
in which RB is called the blockage ratio, describing the percentage of oncoming air flow seen by the building block. The value of RB increases with increasing block height h and decreasing street width LA (simplified as L in the following sections), while the windward length of blocks has multifarious effects on the blockage. The relation between the blockage ratio and the average wind speed along streets is examined in the next section.
T. Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
519
1.0 ~Wtml 0.8
w
o 0.6
~
0.4
0. 2
0.0
W 100 DIO0
o -
50 50 zx 50 100
I
I
,5
10
25 15 25
1
I
L
I
15 20 h (rn)
25
30
Fig. 4. Averaged wind speeds over the along-wind streets between uniform blocks.
1.0
o \
0.8
"~
o o
Murakami Jamieson o
o 0.6
,~
e t al, 1979 o e t al, 1992
Wiren, 1991 a
present
Study •
-
0.4
-
0.2 0.0 0.0
_
/
Us/Uo=I_O.9Rn o-4 I
I
I
I
I
I
0.1
0.2
0.3
0.4
0.5
0.6
0.7
RB
Fig. 5. Empirical relation between the average wind speed and the blockage ratio.
4. W i n d s p e e d o n a l o n g - w i n d streets
An average wind speed /.Ts is taken over all measured points on an along-wind street, as sketched in Fig. 4, excluding 14 points on cross-wind streets. The value of Us decreases with increasing block height, whereas the street width and the plan size of the blocks also show their impact on the average wind speed. The integrated impact can be described as a function of the blockage ratio RB by an empirical relation derived through data fitting,
Us/Uo =
1 - 0.9R °'4 .
(3I
The agreement of such a relation with the wind-tunnel results is demonstrated in Fig. 5 containing data from Refs. 13-5] and the present study. According to this relation, wind speeds along streets decrease with increasing blockage ratios. For
520
T. Stathopoulos, H. Wu/~ WindEng. lnd. Aerodyn. 54/55 (1995) 515-525 U/Uo ~. W
.
iil
h
I
.25 .32 .39
D
.25 ,34 ,39 .33
i!i!i i
.35
L
.40
'I:i •.
.73
.34 . 33
loll
L-25m
.41
i
.78 .72 . ,
- -
L
(a)
U/Uo
U/Us- I for H / h - I
.73
;~;
: •:
.67 .72
Wind
(b)
i
U/Us
U/Uo
!
U/Us
i ,48 .44 ,22
i
.57 .69 .80
.43
! I • 69 .37 .15 ~ ',! . 28 4"/ .ill i 4.0 80 1.2 1.3 .82;2.9 5.0 .6280 .60 ! 1.6 1.1[:::'.'.'.'.'.j.'.'.'.'.'.'.12.3
fi,'.-'" ~ i"" ~ i i": •: i"l "., : : ~ " ' i","
. "34 1'1
•
.21 .40 .21
•
.60 .56 1 . 1 -[ 2.1 3.2 1.7 1.5 t.5
i
• 6,.-
,
.~o
1 . o ....
1,8
1.1
1.3
• 2~
.23
3.6
3.0
.9,1
.89
[
,2
.el1 . ::::::::::::::::::::::::::::::::4 . 93 . 5 1
.57
.34
.
.21:1 .87 .62 .37~'~.3 2.0 .45 I, I .3
.78
I !
.47
,
.f15 1.2 .50
. 2 . - -
,I
.34 1.2 ,8(] .41!1,4 1.9 1.7 1.1 .40 .6B ! 2.0 .S? .47 1.1 .97 ] 2.0 1.5 .83| --.27
I
(c)
.78 .46 ,69
i ;II I
L
1.2 1,2 .63 ,35 - -
i
(d)
Fig. 6. Wind-speed ratios distributed around buildings standing among uniform street blocks.
example, suburban terrains with low RR values cannot provide proper shelter for street winds, whereas wind flows may skim over downtown cores of large cities with high RB values without disturbing pedestrians on streets. However, this is true only for street blocks with the same, or comparable, heights. If a building is much taller than its surroundings, wind patterns on streets around the building m a y change significantly.
5. Tall building above uniform surroundings
The street-level wind patterns become different when the height of the main building increases, as indicated in Fig. 6 - note that the building models in Figs. 6 b - 6 d are not plotted to scale. The m a x i m u m values of U/Uo are 0.74, 1.3 and 1.7, and those of U / U s are 1.0, 5.0 and 5.7 for H / h = 1, 2 and 4, respectively.
T. Stathopoulos, IK Wu/,L Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
521
2.0 I
O
0 1.6
o
~
o
[]
L(m) h(m) 25 15
-~1.2 0.8
9.5 19.1
n
• •
28.6
o
•
[]
b = 2m
0.4
I
o
I
10
I
I
I
I
20 30 (H-h)/b
I
1
40
J
50
Fig. 7. Variation of maximum wind-speedamplificationwith height difference.
The speed variation may be attributed to several geometric factors such as H, h, W and L, but the most direct cause is the height difference (H - h). For a tall building in a built-up region, only the building surface above surroundings obstructs and deflects air flow down to the street level. For a number of building configurations, the maximum mean speeds (UM/Uo) increase with the height difference as shown in Fig. 7. Although the selection of measurement height (b) to non-dimensionalize the height difference seems unusual, it can be justified by the fact that the higher the measurement level, the lower the mean speed ratios such as UM/Uo.Further analysis indicates that the wind speeds also increase with decreasing blockage ratio of surroundings, or increasing average speed over the along-wind street. In other words, the increase of kinetic energy of the wind flow at street level (UM/Uo)stems from the approaching wind flows at both the street level (Us/U0) and higher levels; the latter may relate to the non-dimensional height difference ( H - h)/b. A tentative relation is given as follows, (uM/Uo) ~ = C~(Gs/Uo) 2 + C2
(4)
where b is the measurement height, and C1, C2 and C 3 a r e constants to be determined by experimental data. By using the linear regression technique, constants (C1, C2, C3) are calculated as (2.0, 0.12, 0.88) for data of the present study. These constants are found also applicable for other cases [3-5] when the measurement heights (ranging from2 to 10 m) are taken into account in the wind-speed prediction. The agreement of predicted speeds with measured data is shown in Fig. 8. A more accurate empirical relation may be expected if a wider range of wind-tunnel results are obtained and analyzed. While the highest wind speed normally takes place around the windward corners of a tall building, the vortex flow in front of the tall building is also critical in some building cases. The speed ratio UF/Uo at the central point in front of the main buildings in Fig. 6 is 0.33, 0.82 and 0.71 for H/h = 1, 2 and 4, respectively. More
522
T. Stathopoulos, H. Wu/J. grind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
2.0
0 Murakami
e t al, 1979
._
/"
n Jamieson et al, 1992 ! /• zx Wlren, 1991 O n ,.~/m O • • Current Study mI~ / O ~~ -m
1.5
1.0 q~
m ~ 0.5
a
0.0 O.O
I 0.5
L
[ 1.O
n
I 1.5
n
2.0
Predicted Speeds UM/Uo Fig. 8. Comparison of measured and predicted wind speeds around tall buildings with uniform surroundings.
1.O ~
"o
0.8
[]
~
A
o 0.6
o
?wq
L(m) 25
0.4
h(m) 9.5 19.1
A
•
0.2
28.6
o
•
t
0
[
I0
l
t I I 20 30 (H-h)/b
I
io
15
[] •
b = 2rn 0.0
wind
W
L
40
L
n
50
Fig. 9. Variation of wind speeds in front of tall buildings with height difference.
wind-tunnel data are summarized in Fig. 9 for the speeds in front of tall buildings versus the height difference between buildings. The speed ratio jumps from its lowest value at H - h = 0 to the maximum at H - h = h to 2h. After reaching the maximum, the speed decreases slightly for any additional increase in the height difference for all tested cases. One possible explanation to this phenomenon is that as the height difference becomes larger, the high-level air flow tends to move laterally around the main building, instead of vertically down to the street level. In addition, the intensity, size and position of front vortices depend not only on the height difference, but also on the distance between buildings. Further details on the front vortices between two buildings in tandem positions can be found in Ref. I-2].
T. Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
523
6. Other considerations For all building models discussed so far, the wind flow is parallel, or perpendicular, to streets. Actual winds, however, may come from any direction. For oblique wind directions, the blockage ratio, hence the wind speed, is expected to change. Take uniform street blocks with h = H = 19 m and L = 25 m for example. When W = 100m and D = 50m, wind speeds on along-wind streets decrease from about 0.7 to 0.4 as the wind direction changes from 0 ° to 22.5 ° Under the same circumstance, wind speeds on cross-wind streets do not change much, except at the centre of blocks where wind speeds are more than doubled. When the wind angle becomes 45 °, wind speeds on both streets are higher than those for 22.5 ° or even for 0 ° over some zones. This is also true for the street arrangement with W = 50 m and D = 100 m, for which oblique winds (22.5 ° and 45 °) provide higher speeds in general. This appears contradictory to the well-known guidelines that suggest street blocks be oriented diagonally to prevailing winds for the best protection of pedestrians. However, this finding is limited to uniform blocks with buildings of relatively low heights. More measurements are required to study the speed variation with the incident angle for different configurations of street blocks. In built-up cities, a tall building, instead of possessing the entire block, may be located at the corner or midway of a block. Consequently, its impact on the wind environment may be different on nearby streets I-6,7]. Fig. 10 provides two examples of wind speed distribution along the central line of an adjacent street when the tested tall building has different dimensions and locations, but h = 9.5 m and L = 25 m for all cases. The main building tested is equally divided into two parts with heights H1 and H2. When the two parts have the same height H1/h = H2/h = 8, i.e. a single building covering the entire block, both speed ratios get values higher than other cases. More specifically, (a) If the tall building with a half width of the block is located nearer the street where the measurements occur (H~/h = 8 and H2/h = 1), the speed ratios decrease slightly in comparison to the case ofH~/h = H2/h = 8. When such a building is moved to the far side of the block (H1/h = 1 and H2/h = 8), the maximum values of two speed ratios drop significantly with (U/Uo, U/Us) equal to (1.7, 2.3), (1.7, 2.0) and (1.1, 1.3) for the three cases, respectively. (b) When the windward width W = 50 m, the slender building can be moved along the wind direction within the same block. If the slender building covers the front half of the block (H1/h = 8 and H2/h = 1), it induces high wind speeds on the street similar to those with the tall building covering the entire block. On the other hand, when the slender building is drawn backwards (H1/h = 1 and H2/h = 8), wind speeds are reduced by a noticeable amount (from 1.6 to 1.3 for U/Uo). The location where the maximum speed occurs is also moved downstream accordingly. The speed drops in both cases demonstrate the efficiency on wind reduction of a setback or a large podium in the design of high-rise buildings I-2,6].
T. Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
524 2.5
I'I ~,
2.0 ~
1.5
..1-
Solid Lines: U/UQ Dashed Lines: U/U s
IiiI
-
8 8
• • •
x \
8 1 8
1
W'lO0 I)-50 L-25
L
D
1.0
~0-"
;X
L
0.5 h
s
l
h
0.0 0
50
100
150
200
x (m)
|
.~
~Dashed Lines: U/U~
Ht/h ~ •
• • /
0.5
.,.
=
o.oF 0
50
,
100
150 200 X (m)
I 250
.
.d
8 8
1
h-9.5(m) 8 l
8
W-50 D-100 L-25
D
? 300
350
Fig. 10. Effect of tallbuilding locations within a block on wind speeds along adjacent streets.
7. Conclusions
Wind-tunnel experiments have been carried out for pedestrian-level wind speeds around street blocks. Generic models and empirical relations have been established by combining the current findings with literature information for the wind conditions over streets. They can be expanded for some preliminary wind assessment in the urban environment for specific cases. The conclusions of this paper can be summarized as follows: (1) For uniformly distributed street blocks, a blockage ratio is defined by the windward dimensions of buildings and streets (Eq. (2)). The average wind speed over along-wind streets can be estimated by such a blockage ratio (Eq. (3)). (2) When a tall building is surrounded by uniform blocks, the maximum speed amplification is a function of the height difference of the tall building and the surroundings, the measurement height and the blockage ratio of surrounding blocks. Empirical equations have been suggested for the assessment of wind conditions for generic street cases (Eq. (4)).
T Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525 (3)
525
F u r t h e r e x p e r i m e n t s h a v e b e e n c a r r i e d o u t for the effects of w i n d d i r e c t i o n a n d l o c a t i o n of the tall b u i l d i n g w i t h i n a b l o c k , b u t o n l y g e n e r a l g u i d e l i n e s are p r o v i d e d i n the p a p e r .
References [1] J. Gandemer, Wind environment around buildings: aerodynamic concepts, in: Proc. 4th Int. Conf. on Wind effects on buildings and structures, Heathrow, UK (1975) pp. 423-432. 1-2] A.D. Penwarden and A.F.E. Wise, Wind environment around buildings, Building Research Establishment Report, Department of Environment, Building Research Establishment, H.M.S.O., London, UK, 1975. [3] S. Murakami, K. Uehara and H. Komine, Amplification of wind speed at ground level due to construction of high-rise building in urban area, J. Ind. Aerodyn. 4 (1979) 343-370. 1-4] N.J. Jamieson, P. Carpenter and P.D. Cenek, The effect of architectural detailing on pedestrian-level wind speeds, J. Wind Eng. Ind. Aerodyn. 41-44 (1992) 2301-2312. I-5] B. Wiren, A wind tunnel study of wind speeds near the ground in a group of block-type buildings, Research Report of the National Swedish Institue for Building Research, Giivle, Sweden, 1991. [6] N. Isyumov, S. Helliwell, S. Rosen and D. Lai, Winds in cities: effects on pedestrians and the dispersion of ground level pollutants, in: Proc. 6th Colloq. on Industrial Aerodynamics, Aachen, Germany, 1985. [7] F.M. Livesey, The impact of new development on pedestrian-level winds, M.E.Sc. thesis, University of Western Ontario, London, Ontario, Canada, 1991. [8] H.P.A.H. Irwin, A simple omnidirectional sensor for wind-tunnel studies of pedestrian-level winds, J. Wind Eng. Ind. Aerodyn. 7 (1981) 219-239. [9] H. Wu and T. Stathopoulos, Further experiments on Irwin's surface wind sensor, J. Wind Eng. Ind. Aerodyn. 53 (1994) 441-452. [10] T.R. Oke, Street design and urban canopy layer climate, Energy Build. 11 (1988) 103 111.
ELSEVIER
Journal of Wind Engineering and Industrial Aerodynamics54/55 (1995) 515-525
~/~1/~
Generic models for pedestrian-level winds in built-up regions Theodore Stathopoulos*, Hanqing Wu Centre for Building Studies, Concordia University, 1455 de Maisonneuve Blvd. W., Montreal Quebec, Canada H3G 1M8
Abstract
Wind conditions over streets in built-up cities were examined in a boundary-layer windtunnel model study. Presented in this paper are results of wind speeds affected by a number of parameters such as the spatial density of street blocks, the building height over surroundings, the relative location of buildings and the direction of approaching wind. Based on literature information and current findings, some generic models have been established with the empirical relations between wind conditions and building configurations. These models can be applied for the preliminary estimation of pedestrian-level wind environmental conditions in built-up regions.
1. Introduction
Wind conditions over streets are an increasing concern for pedestrian comfort, pollutant dispersion, snow accumulation, etc. Such conditions depend on meteorological data, upstream terrains, surrounding blocks and adjacent buildings under consideration. Wind-tunnel experiments have been carried out for simplified models such as isolated buildings, twin buildings and more complex building combinations [1, 2-1. For several simple models, parametric relations have been attempted between wind speeds and building configurations including dimensions, geometries and relative positions. However, those relations cannot directly apply for practical situations unless the influence of surroundings and other impact parameters is properly taken into consideration. Extensive wind-tunnel experiments were performed by Murakami et al. [3] for tall buildings surrounded by uniformly distributed square blocks. The impact on wind flows was investigated with respect to building height, street width, surrounding height, wind direction and other parameters. A similar experiment was reported by
*Corresponding author. 0167-6105/95/$09.50 © 1995 ElsevierScienceB.V. All rights reserved SSD1 0 1 6 7 - 6 1 0 5 ( 9 4 ) 0 0 0 6 8 - O
516
T. Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
Jamieson et al. [4], concentrating on the effect of architectural details of the tall buildings. The Wiren study [5] was carried out for a group of nine buildings with average and maximum wind speeds presented as functions of wind directions. In studies reported by Isyumov et al. [6] and Livesey [7], street blocks were constituted of several buildings of different heights so that the distribution and the orientation of large buildings within the block became influential to wind patterns over adjacent streets. In order to generalize the knowledge of wind conditions over streets in built-up cities, research has been carried out on (1) the measurement of wind speeds along streets between uniform blocks; (2) the evaluation of wind effects of tall buildings surrounded by uniform street blocks; and (3) the analysis and synthesis of the results obtained from generic models which simulate actual city configurations. This paper presents details of the wind-tunnel experiments and the generic models for wind conditions over streets. Although the application of the established models may be limited at the present stage, these generic models and associated empirical relations are a useful progress towards some preliminary assessment of wind conditions in the urban environment.
2. Experimental setup Experiments were conducted in a boundary-layer wind tunnel designed for building aerodynamics applications. Suburban wind conditions were simulated with a mean speed profile of power-law exponent equal to 0.25 and a geometric scale of 1:500. Building models were made of wood pieces with the same dimensions of 19x 1 0 0 x 2 0 0 m m , corresponding to full-scale street blocks 9.5m high and 50 x 100 m in plan size. The block height could be increased easily by piling up wood pieces, as shown in Fig. 1. In a typical model arrangement, there were usually five rows of blocks upstream of the main building which was located at the centre of the turntable of wind tunnel. In the current study, both the block height h and the main building height H were multiples of 9.5 m, the plan dimensions W and D were either 100 m or 50 m depending on the wind direction, the width of along-wind streets (LA) was 25 m or 15 m whereas that of cross-wind streets (Lc) was kept constant at 25 m. Irwin's surface wind sensors [8] were utilized considering their advantages in measuring a large number of horizontal wind speeds - see also the additional work carried out on these sensors [9]. With the height of 4 mm (2 m in full scale) above the ground, 37 sensors were distributed as shown in Fig. 2 for street widths of 25 m. The least distance between sensors was 20 mm in order to avoid any possible interference. The sensors were installed in the centre of the streets and at distances of 5 mm (2.5 m in full scale) from buildings, mainly on sidewalks and crosswalks of streets where both high-speed winds and pedestrian activities usually take place. While most attention was paid to the along-wind street, the wind speeds on cross streets were also measured on crosswalks, as well as in front of and behind the main building. Such a sensor arrangement also applied for the street models with the short side (50 m) facing the oncoming wind. When a narrow street (LA = 15 m) was tested, six measurement points would be covered by building models.
T Stathopoulos, H. Wu/J. WindEng. Ind. Aerodyn. 54/55 (1995) 515-525
517
D
Wind
LA
~~L c D
Fig. 1. A typicalmodel arrangementwith a tall buildingsurrounded by uniformstreet blocks.
Lc
ooI
D
•
W
Fig. 2. Distribution of measurementpoints on streets 25 m wide.
In this study, mean wind speeds (U) are normalized by two reference speeds, namely, U0, the mean wind speed at each measured point when no building models are present, and Us, the wind speed on streets between uniform blocks when all street blocks are of the same height (h = H). By definition, U/Us = 1 at all measured points when H/h = 1. For the same building configuration, values of U/Uo are usually lower than one due to the sheltering effect caused by buildings.
3. Blockage ratio
For uniform street blocks, wind speeds along streets are mainly dependent upon the distribution and density of street blocks. If buildings on a street block are typified by a single cuboid as shown in Fig. 3, the following two area ratios can be defined according to Oke [10]: (1) the plan density, Art/AL, which is the ratio of the roof area
518
T. Stathopoulos, H. Wu/J. WindEng. Ind. Aerodyn. 54/55 (1995) 515-525
AF i
D
:
Fig. 3. A sketch of simplifiedstreet blocks: dimensionsand areas.
of building over that of the building lot; and (2) the roughness density, As/AL, based on the front area of the building. These ratios can be expressed as follows,
AR
WD
As
Wh
AL
(W + LA) (D + Lc)'
AL
(W + LA) (D + Lc)'
(1)
where W and D are cross- and along-wind lengths of a block, Lc and LA are widths of cross- and along-wind streets around the block, and h is the block height. The two ratios have the same value if the street block is a cube, i.e. W = D = h. The roughness density is related not only to the plan dimensions, but also to the building height; it may be more indicative than the plan density from the viewpoint of wind conditions on streets. However, both ratios are based on the lot area, that, sometimes, may not be quite suitable for the representation of street densities and wind intensities. For non-cubic blocks, an increase in the along-wind length (D) will, by definition, reduce the roughness density, but may increase the plan density in certain cases. On the other hand, it is well known that the effect of along-wind length is insignificant on the wind speeds around building corners and between two side-by-side buildings. It is windward building surfaces that affect the oncoming wind flow and influence the wind conditions at street level. Based on these considerations, a new area ratio using only windward areas of buildings and streets is proposed here, As
R~ - AF
Wh
(W + LA) 2 '
(2)
in which RB is called the blockage ratio, describing the percentage of oncoming air flow seen by the building block. The value of RB increases with increasing block height h and decreasing street width LA (simplified as L in the following sections), while the windward length of blocks has multifarious effects on the blockage. The relation between the blockage ratio and the average wind speed along streets is examined in the next section.
T. Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
519
1.0 ~Wtml 0.8
w
o 0.6
~
0.4
0. 2
0.0
W 100 DIO0
o -
50 50 zx 50 100
I
I
,5
10
25 15 25
1
I
L
I
15 20 h (rn)
25
30
Fig. 4. Averaged wind speeds over the along-wind streets between uniform blocks.
1.0
o \
0.8
"~
o o
Murakami Jamieson o
o 0.6
,~
e t al, 1979 o e t al, 1992
Wiren, 1991 a
present
Study •
-
0.4
-
0.2 0.0 0.0
_
/
Us/Uo=I_O.9Rn o-4 I
I
I
I
I
I
0.1
0.2
0.3
0.4
0.5
0.6
0.7
RB
Fig. 5. Empirical relation between the average wind speed and the blockage ratio.
4. W i n d s p e e d o n a l o n g - w i n d streets
An average wind speed /.Ts is taken over all measured points on an along-wind street, as sketched in Fig. 4, excluding 14 points on cross-wind streets. The value of Us decreases with increasing block height, whereas the street width and the plan size of the blocks also show their impact on the average wind speed. The integrated impact can be described as a function of the blockage ratio RB by an empirical relation derived through data fitting,
Us/Uo =
1 - 0.9R °'4 .
(3I
The agreement of such a relation with the wind-tunnel results is demonstrated in Fig. 5 containing data from Refs. 13-5] and the present study. According to this relation, wind speeds along streets decrease with increasing blockage ratios. For
520
T. Stathopoulos, H. Wu/~ WindEng. lnd. Aerodyn. 54/55 (1995) 515-525 U/Uo ~. W
.
iil
h
I
.25 .32 .39
D
.25 ,34 ,39 .33
i!i!i i
.35
L
.40
'I:i •.
.73
.34 . 33
loll
L-25m
.41
i
.78 .72 . ,
- -
L
(a)
U/Uo
U/Us- I for H / h - I
.73
;~;
: •:
.67 .72
Wind
(b)
i
U/Us
U/Uo
!
U/Us
i ,48 .44 ,22
i
.57 .69 .80
.43
! I • 69 .37 .15 ~ ',! . 28 4"/ .ill i 4.0 80 1.2 1.3 .82;2.9 5.0 .6280 .60 ! 1.6 1.1[:::'.'.'.'.'.j.'.'.'.'.'.'.12.3
fi,'.-'" ~ i"" ~ i i": •: i"l "., : : ~ " ' i","
. "34 1'1
•
.21 .40 .21
•
.60 .56 1 . 1 -[ 2.1 3.2 1.7 1.5 t.5
i
• 6,.-
,
.~o
1 . o ....
1,8
1.1
1.3
• 2~
.23
3.6
3.0
.9,1
.89
[
,2
.el1 . ::::::::::::::::::::::::::::::::4 . 93 . 5 1
.57
.34
.
.21:1 .87 .62 .37~'~.3 2.0 .45 I, I .3
.78
I !
.47
,
.f15 1.2 .50
. 2 . - -
,I
.34 1.2 ,8(] .41!1,4 1.9 1.7 1.1 .40 .6B ! 2.0 .S? .47 1.1 .97 ] 2.0 1.5 .83| --.27
I
(c)
.78 .46 ,69
i ;II I
L
1.2 1,2 .63 ,35 - -
i
(d)
Fig. 6. Wind-speed ratios distributed around buildings standing among uniform street blocks.
example, suburban terrains with low RR values cannot provide proper shelter for street winds, whereas wind flows may skim over downtown cores of large cities with high RB values without disturbing pedestrians on streets. However, this is true only for street blocks with the same, or comparable, heights. If a building is much taller than its surroundings, wind patterns on streets around the building m a y change significantly.
5. Tall building above uniform surroundings
The street-level wind patterns become different when the height of the main building increases, as indicated in Fig. 6 - note that the building models in Figs. 6 b - 6 d are not plotted to scale. The m a x i m u m values of U/Uo are 0.74, 1.3 and 1.7, and those of U / U s are 1.0, 5.0 and 5.7 for H / h = 1, 2 and 4, respectively.
T. Stathopoulos, IK Wu/,L Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
521
2.0 I
O
0 1.6
o
~
o
[]
L(m) h(m) 25 15
-~1.2 0.8
9.5 19.1
n
• •
28.6
o
•
[]
b = 2m
0.4
I
o
I
10
I
I
I
I
20 30 (H-h)/b
I
1
40
J
50
Fig. 7. Variation of maximum wind-speedamplificationwith height difference.
The speed variation may be attributed to several geometric factors such as H, h, W and L, but the most direct cause is the height difference (H - h). For a tall building in a built-up region, only the building surface above surroundings obstructs and deflects air flow down to the street level. For a number of building configurations, the maximum mean speeds (UM/Uo) increase with the height difference as shown in Fig. 7. Although the selection of measurement height (b) to non-dimensionalize the height difference seems unusual, it can be justified by the fact that the higher the measurement level, the lower the mean speed ratios such as UM/Uo.Further analysis indicates that the wind speeds also increase with decreasing blockage ratio of surroundings, or increasing average speed over the along-wind street. In other words, the increase of kinetic energy of the wind flow at street level (UM/Uo)stems from the approaching wind flows at both the street level (Us/U0) and higher levels; the latter may relate to the non-dimensional height difference ( H - h)/b. A tentative relation is given as follows, (uM/Uo) ~ = C~(Gs/Uo) 2 + C2
(4)
where b is the measurement height, and C1, C2 and C 3 a r e constants to be determined by experimental data. By using the linear regression technique, constants (C1, C2, C3) are calculated as (2.0, 0.12, 0.88) for data of the present study. These constants are found also applicable for other cases [3-5] when the measurement heights (ranging from2 to 10 m) are taken into account in the wind-speed prediction. The agreement of predicted speeds with measured data is shown in Fig. 8. A more accurate empirical relation may be expected if a wider range of wind-tunnel results are obtained and analyzed. While the highest wind speed normally takes place around the windward corners of a tall building, the vortex flow in front of the tall building is also critical in some building cases. The speed ratio UF/Uo at the central point in front of the main buildings in Fig. 6 is 0.33, 0.82 and 0.71 for H/h = 1, 2 and 4, respectively. More
522
T. Stathopoulos, H. Wu/J. grind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
2.0
0 Murakami
e t al, 1979
._
/"
n Jamieson et al, 1992 ! /• zx Wlren, 1991 O n ,.~/m O • • Current Study mI~ / O ~~ -m
1.5
1.0 q~
m ~ 0.5
a
0.0 O.O
I 0.5
L
[ 1.O
n
I 1.5
n
2.0
Predicted Speeds UM/Uo Fig. 8. Comparison of measured and predicted wind speeds around tall buildings with uniform surroundings.
1.O ~
"o
0.8
[]
~
A
o 0.6
o
?wq
L(m) 25
0.4
h(m) 9.5 19.1
A
•
0.2
28.6
o
•
t
0
[
I0
l
t I I 20 30 (H-h)/b
I
io
15
[] •
b = 2rn 0.0
wind
W
L
40
L
n
50
Fig. 9. Variation of wind speeds in front of tall buildings with height difference.
wind-tunnel data are summarized in Fig. 9 for the speeds in front of tall buildings versus the height difference between buildings. The speed ratio jumps from its lowest value at H - h = 0 to the maximum at H - h = h to 2h. After reaching the maximum, the speed decreases slightly for any additional increase in the height difference for all tested cases. One possible explanation to this phenomenon is that as the height difference becomes larger, the high-level air flow tends to move laterally around the main building, instead of vertically down to the street level. In addition, the intensity, size and position of front vortices depend not only on the height difference, but also on the distance between buildings. Further details on the front vortices between two buildings in tandem positions can be found in Ref. I-2].
T. Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
523
6. Other considerations For all building models discussed so far, the wind flow is parallel, or perpendicular, to streets. Actual winds, however, may come from any direction. For oblique wind directions, the blockage ratio, hence the wind speed, is expected to change. Take uniform street blocks with h = H = 19 m and L = 25 m for example. When W = 100m and D = 50m, wind speeds on along-wind streets decrease from about 0.7 to 0.4 as the wind direction changes from 0 ° to 22.5 ° Under the same circumstance, wind speeds on cross-wind streets do not change much, except at the centre of blocks where wind speeds are more than doubled. When the wind angle becomes 45 °, wind speeds on both streets are higher than those for 22.5 ° or even for 0 ° over some zones. This is also true for the street arrangement with W = 50 m and D = 100 m, for which oblique winds (22.5 ° and 45 °) provide higher speeds in general. This appears contradictory to the well-known guidelines that suggest street blocks be oriented diagonally to prevailing winds for the best protection of pedestrians. However, this finding is limited to uniform blocks with buildings of relatively low heights. More measurements are required to study the speed variation with the incident angle for different configurations of street blocks. In built-up cities, a tall building, instead of possessing the entire block, may be located at the corner or midway of a block. Consequently, its impact on the wind environment may be different on nearby streets I-6,7]. Fig. 10 provides two examples of wind speed distribution along the central line of an adjacent street when the tested tall building has different dimensions and locations, but h = 9.5 m and L = 25 m for all cases. The main building tested is equally divided into two parts with heights H1 and H2. When the two parts have the same height H1/h = H2/h = 8, i.e. a single building covering the entire block, both speed ratios get values higher than other cases. More specifically, (a) If the tall building with a half width of the block is located nearer the street where the measurements occur (H~/h = 8 and H2/h = 1), the speed ratios decrease slightly in comparison to the case ofH~/h = H2/h = 8. When such a building is moved to the far side of the block (H1/h = 1 and H2/h = 8), the maximum values of two speed ratios drop significantly with (U/Uo, U/Us) equal to (1.7, 2.3), (1.7, 2.0) and (1.1, 1.3) for the three cases, respectively. (b) When the windward width W = 50 m, the slender building can be moved along the wind direction within the same block. If the slender building covers the front half of the block (H1/h = 8 and H2/h = 1), it induces high wind speeds on the street similar to those with the tall building covering the entire block. On the other hand, when the slender building is drawn backwards (H1/h = 1 and H2/h = 8), wind speeds are reduced by a noticeable amount (from 1.6 to 1.3 for U/Uo). The location where the maximum speed occurs is also moved downstream accordingly. The speed drops in both cases demonstrate the efficiency on wind reduction of a setback or a large podium in the design of high-rise buildings I-2,6].
T. Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525
524 2.5
I'I ~,
2.0 ~
1.5
..1-
Solid Lines: U/UQ Dashed Lines: U/U s
IiiI
-
8 8
• • •
x \
8 1 8
1
W'lO0 I)-50 L-25
L
D
1.0
~0-"
;X
L
0.5 h
s
l
h
0.0 0
50
100
150
200
x (m)
|
.~
~Dashed Lines: U/U~
Ht/h ~ •
• • /
0.5
.,.
=
o.oF 0
50
,
100
150 200 X (m)
I 250
.
.d
8 8
1
h-9.5(m) 8 l
8
W-50 D-100 L-25
D
? 300
350
Fig. 10. Effect of tallbuilding locations within a block on wind speeds along adjacent streets.
7. Conclusions
Wind-tunnel experiments have been carried out for pedestrian-level wind speeds around street blocks. Generic models and empirical relations have been established by combining the current findings with literature information for the wind conditions over streets. They can be expanded for some preliminary wind assessment in the urban environment for specific cases. The conclusions of this paper can be summarized as follows: (1) For uniformly distributed street blocks, a blockage ratio is defined by the windward dimensions of buildings and streets (Eq. (2)). The average wind speed over along-wind streets can be estimated by such a blockage ratio (Eq. (3)). (2) When a tall building is surrounded by uniform blocks, the maximum speed amplification is a function of the height difference of the tall building and the surroundings, the measurement height and the blockage ratio of surrounding blocks. Empirical equations have been suggested for the assessment of wind conditions for generic street cases (Eq. (4)).
T Stathopoulos, H. Wu/J. Wind Eng. Ind. Aerodyn. 54/55 (1995) 515-525 (3)
525
F u r t h e r e x p e r i m e n t s h a v e b e e n c a r r i e d o u t for the effects of w i n d d i r e c t i o n a n d l o c a t i o n of the tall b u i l d i n g w i t h i n a b l o c k , b u t o n l y g e n e r a l g u i d e l i n e s are p r o v i d e d i n the p a p e r .
References [1] J. Gandemer, Wind environment around buildings: aerodynamic concepts, in: Proc. 4th Int. Conf. on Wind effects on buildings and structures, Heathrow, UK (1975) pp. 423-432. 1-2] A.D. Penwarden and A.F.E. Wise, Wind environment around buildings, Building Research Establishment Report, Department of Environment, Building Research Establishment, H.M.S.O., London, UK, 1975. [3] S. Murakami, K. Uehara and H. Komine, Amplification of wind speed at ground level due to construction of high-rise building in urban area, J. Ind. Aerodyn. 4 (1979) 343-370. 1-4] N.J. Jamieson, P. Carpenter and P.D. Cenek, The effect of architectural detailing on pedestrian-level wind speeds, J. Wind Eng. Ind. Aerodyn. 41-44 (1992) 2301-2312. I-5] B. Wiren, A wind tunnel study of wind speeds near the ground in a group of block-type buildings, Research Report of the National Swedish Institue for Building Research, Giivle, Sweden, 1991. [6] N. Isyumov, S. Helliwell, S. Rosen and D. Lai, Winds in cities: effects on pedestrians and the dispersion of ground level pollutants, in: Proc. 6th Colloq. on Industrial Aerodynamics, Aachen, Germany, 1985. [7] F.M. Livesey, The impact of new development on pedestrian-level winds, M.E.Sc. thesis, University of Western Ontario, London, Ontario, Canada, 1991. [8] H.P.A.H. Irwin, A simple omnidirectional sensor for wind-tunnel studies of pedestrian-level winds, J. Wind Eng. Ind. Aerodyn. 7 (1981) 219-239. [9] H. Wu and T. Stathopoulos, Further experiments on Irwin's surface wind sensor, J. Wind Eng. Ind. Aerodyn. 53 (1994) 441-452. [10] T.R. Oke, Street design and urban canopy layer climate, Energy Build. 11 (1988) 103 111.