# Probability analysis of wind-driven natural ventilation potential in areaway-attached basements.

IntroductionIn recent decades, with ongoing urbanization and increasing needs for more living space in urban residential areas, it has become preferable to design a basement as an extra space for residential purposes in a dwelling house, especially in some crowded metropolitan cities. However, despite many acclaimed merits and benefits of subterranean dwelling in energy-efficient aspects (Al-Mumin 2001), there are still some environmental constraints, among which the air quality in basements is the most concerning. There are a variety of ways to meet the air quality requirement, either through mechanical systems or via natural forces, and there is greater need for some people to ventilate than to just meet air quality requirements (Russell et al. 2007). As an energy-effective solution and also a feature of sustainable building design, natural ventilation has the potential to maintain an acceptable indoor environment without the necessity of mechanical means, if it is well integrated into building systems through openings and louvers, etc. Nevertheless, the achievement of natural ventilation in basements is not as easy as in above-ground rooms. In urban areas of Japan, attempts have been made to introduce natural ventilation into basement rooms for residential purposes by using so-called areaway space. As shown in Figure 1, areaway space, as it is referred to throughout this article, is considered to be an open sunken courtyard for the generation of private space, allowing access, air, and light to its adjacent basement through openings between them. The presence of such a space can act as a buffer zone to moderate the local microclimate in the adjacent basement room.

[FIGURE 1 OMITTED]

From the perspective of indoor-air quality, the primary benefit of an areaway is the improved natural ventilation performances in basement spaces. This article presents a study of natural ventilation potential in areaway-attached basements. Since cross-ventilation is restricted in most cases, and the stack effect is also limited due to lack of space, only single-sided ventilation is considered for areaway attached basements in urban dwellings. Namely, the attention of this study is focused on single-sided natural ventilation driven by wind force. A few studies have been conducted concerning wind-induced natural ventilation in areaway-attached basements or similar architectural forms, such as underground dwellings or courtyards. Hall et al. (1999) conducted a wind-tunnel experiment to study the dispersion of contaminants in courtyards, and a wide range of courtyard parameters, including ratio of depth to width, wind direction, wall thickness, etc., were investigated using small-scale models. Rajapaksha et al. (2003) investigated the potential of an internal courtyard to minimize indoor overheating conditions in a single-story building located in a warm humid climate using field measurements and computational fluid dynamics (CFD) simulations. Recently, Bu et al. (2010) carried out a wind-tunnel experiment on an areaway-attached model exposed to an urban atmospheric boundary layer. The experimental results indicated that the effective ventilation rate in the basement depended greatly on ambient wind direction. Some other studies on single-sided natural ventilation also found the significant impact of ambient wind direction on indoor ventilation performances (Warren and Parkins 1985; Narasaki et al. 1989; Larsen and Heiselberg 2008).

Most of these studies investigated the effect of one or several wind directions individually based on deterministic approaches (Pietrzyk and Hagentoft 2008). However, urban wind is a stochastic phenomenon, constantly changing in speed and direction. The uncertainties associated with the influences of stochastic wind conditions on indoor ventilation performances can be significant. From this perspective, the analysis of natural ventilation performance should be based on a probabilistic approach rather than on a deterministic approach, which can take stochastic wind conditions into full account. In addition, as wind conditions differ from city to city, it is quite important for building designers and owners to know the probability of ensuring acceptable indoor air quality by natural ventilation and to assess the applicability of areaways at the preliminary design stage of basements.

The present study introduces the concept of exceedance probability analysis based on an air change rate (ACR) index for natural ventilation potential analysis in areaway-attached basements. First, the probabilistic approaches based on different indices will be introduced in the next section, followed by the description of the measurement of ACR index in a wind-tunnel experiment using a two-story building model with an areaway-attached basement. using the measured ACR and local meteorological statistical data, including Weibull parameters and occurrence frequencies of each wind direction, the exceedance probabilities will be calculated for all respective wind directions, and the overall probability distributions are summarized as a function of a series of specified ACR values for each investigated model orientation. Furthermore, a simple evaluation approach for natural ventilation potential analysis will be proposed by extracting ACRs from the calculated exceedance probability distribution curves, which correspond to probabilities of 85% and 15%, respectively. By using the exceedance probability analysis method and the simple evaluation approach, the influence of a variety of parameters, including opening configuration, areaway plan area, building coverage ratio, orientation, and construction site of the model, are further investigated.

Exceedance probability analysis methods

As mentioned above, the urban wind environment is a stochastic phenomenon, and the ventilation performance in a naturally ventilated building changes accordingly. To address the uncertainties associated with urban wind in the analysis of natural ventilation potential, the concept of exceedance probability based on an ACR index is introduced for the analysis of natural ventilation potential in areaway-attached basement spaces.

As a probabilistic approach, the exceedance probability analysis has been applied by some previous researchers to investigate ground-level wind comfort at specific locations in an urban context, particularly around tall buildings. For instance, Murakami et al. (1986) constructed a criterion on the basis of local wind speed at height of 1.5 m for determining the acceptable probability of wind environment in a built-up area in Tokyo. Several similar criteria (Hunt et al. 1976; Melbourne 1978) have been developed and applied in different countries. Although different in implementation, these existing criteria are based on the percent time that mean or peak wind speed at a given location is exceeded annually (Ratcliff and Peterka 1990). In practice, the wind speed at a given location is often assumed to follow a two-parameter Weibull distribution. The Weibull parameters can be derived from those estimated for the wind speed at reference height (usually at nearby meteorological station), based on the assumption of linear correlation between the local wind speed and the wind speed at reference height; e.g., the wind velocity ratios R([a.sub.n]) shown in Equation 1 are unchanged for each azimuth [a.sub.n].

R([a.sub.n]) = Vg/V ([a.sub.n]), (1)

where [V.sub.g] is the local wind speed at ground level, and V is the wind speed at reference height.

The velocity ratios are usually determined by wind-tunnel experiment or CFD simulation. By virtue of occurrence frequencies of wind direction dividedinto16 azimuth sectors, the total exceedance probability for local wind speed at ground level can be summarized and expressed in the following form:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where P(>[V.sub.g,s]) represents the total probability of exceeding a given wind speed [V.sub.g,s] at ground level, A([a.sub.n]) is the relative frequency of occurrence of wind from azimuth an, and K([a.sub.n]) and C([a.sub.n]) are the shape and scale parameters of Weibull distribution for azimuth [a.sub.n].

According to Equation 2, the scale of wind speed is used as the index to calculate the exceedance probabilities in the evaluation of local wind environment at specific locations. Therefore, the corresponding criterion, referred to as velocity-based exceedance probability (Bu et al. 2009), seems more applicable to specific locations than a specific domain, e.g., the basement space in the present study. In order to evaluate local wind environment for a specific domain in terms of natural ventilation performance and thermal comfort, Bu et al. (2009) proposed two new criteria based on two indices: ACR ([h.sup.-1]) and spatially averaged kinetic energy ([m.sup.2]/[s.sup.2])of the target domain, respectively. In light of these two criteria, it is possible to evaluate ventilation capacity and airflow capacity of the whole target domain in a probabilistic approach. Since it is with the interest of natural ventilation potential analysis in areaway-attached basements that this article is concerned, attention is restricted to the first criterion (hereinafter referred to as ACR-based exceedance probability), which is expressed as follows:

P(> [ACR.sub.g,s]) = [15.summation over [n=0] A([a.sub.n])

x exp {-([ACR.sub.g,s]/[R.sub.n]([a.sub.n]) x C([a.sub.n])).sup.K([a.sub.n])}. (3)

The above equation has the similar form as Equation 2, where P(> [ACR.sub.g,s]) represents the total probability of exceeding a given [ACR.sub.g,s] in the target domain. [R.sub.N]([a.sub.n]) is the ratio between the ACR in the target domain and the wind speed at reference height for azimuth [a.sub.n] (see Equation 4), which can be obtained likewise by wind-tunnel experiment or CFD simulation:

[R.sub.N]([a.sub.n]) = ACR/V ([a.sub.n]) = 36000 x PFR/V ([a.sub.n]) x vol, (4)

where purging flow rate (PFR) ([m.sup.3]/s) represents the effective ventilation rate required to remove pollutants from the target domain (Kato et al. 2003), and vol is the domain volume ([m.sup.3]). In the present study, the ACRs in areaway-attached basements were obtained for each wind direction by a wind-tunnel experiment, which will be described in the following section.

Wind-tunnel experiment

The wind-tunnel experiment was performed for areaway-attached basement models in the boundary layer wind tunnel of the Institute of Industrial Science at the University of Tokyo. The wind tunnel has a test section of 16.5 m (54.1 ft) long, 2.2 m (7.2 ft) wide, and 1.8 m (5.9 ft) high, and all measurements were carried out under neutral atmospheric conditions. An approaching wind profile was generated to simulate the atmospheric boundary layer in residential areas, which followed a power law function with an exponent of 0.25. All tests were performed with a free stream wind speed of approximately 3 m/s (9.8 ft/s) and a turbulence intensity of 17% at the building height (0.2 m [0.66 ft]). A more detailed description regarding the wind-tunnel setup and the experimental procedures can be found in previous literature (Bu et al. 2010).

The basement model used in the experiment was in simplified geometries with a scale of 1:30, representing a two-story residential building with an areaway-attached basement. The length and height of the areaway space are fixed, while its width (w) can be changed to allow investigation of the influence of areaway plan area (A) on ventilation performance in the basement space. Figures 2 and 3 show the sketch and photograph of the model with the default configuration, i.e. a single opening of 1.8 x 1.8 [m.sup.2] (5.9 x 5.9 [ft.sup.2]) and an areaway plan area of 18 [m.sup.2] (193.8 [ft.sup.2]) in prototype size.

Figure 4 shows the definitions regarding azimuths direction, orientation of the basement model, wind incidence angle relative to the basement model, as well as their relationships. As depicted in this figure, the horizontal plane is divided into 16 azimuths direction and marked by [a.sub.n], starting from north-northeast (NNE) to north (N) in a clockwise order, i.e., [a.sub.0] = NNE, [a.sub.1] = NE, ... , [a.sub.15] = N. The orientation of the basement model ([beta]) is defined as the outward direction perpendicular to the opening. The approaching wind is also divided into 16 directions marked by i, and the angle between the approaching wind and [beta] is defined as wind incidence angle in a clockwise order and marked by [theta].

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

The determination of the effective ventilation rate, i.e., PFR, requires the measurement of spatially averaged concentration under the conditions of uniformly releasing of tracergas (Buetal. 2010). In this experiment, ethylene was used as the tracer gas, released and sampled at a rate of q = 1.67e -6 [m.sup.3]/s (5.9e-5 [ft.sup.3]/s), with 48 metallic pipes uniformly arranged within the basement space. Given the releasing rate of tracer gas q ([m.sup.3]/s) and the internal volume of the basement space ([m.sup.3]), the ACR (h-1) in the basement can be calculated from the spatial averaged concentration <c> ([m.sup.3]/[m.sup.3])as follows:

ACR = 3600 x PFR/vol = 3600 x q/vol x < c >. (5)

Table 1 shows the parameters investigated using the exceedance probability analysis, including opening configuration, areaway plan area building coverage ratio (R), orientation, and construction site of the model. The dimensions shown in Table 1 are all based on prototype size. During the measurement, the model was rotated with a 22.5[degrees] pitch to measure the concentration in the basement for different wind directions. Considering the symmetry of the model, nine incidence angles were measured between 0[degrees] and 180[degrees] for each case. The building coverage ratio was investigated by a arranging dummy building in a regular array with variable separation distances. R = 0 represents the detached building pattern, while R > 0 represents the multi-building pattern. The exceedance probability results were compared for the basement model between three different sites: Tokyo, Osaka, and Sendai. Figure 5 shows the occurrence frequencies of 16 wind directions in these 3 cities, respectively, which are based on the wind statistical data over 10 years from local meteorological stations. As depicted in this figure, the prevailing wind direction of Tokyo is north-northwest (NNW, 20.6%), followed by N (12.0%), and SW (11.0%), while relatively even distributions can be found for Osaka and Sendai.

[FIGURE 5 OMITTED]

Results and discussion

ACR in the areaway-attached basement

Based on the experimental results of spatially averaged concentration, the ACR in the areaway-attached basement was calculated according to Equation 5 and then converted to the value of the prototype size at a scale of 1:30. The following discussions are all based on the results of prototype size. Among the parameters listed in Table 1, opening configuration, areaway plan area, and building coverage ratio may have influences on the ACR in the basement, while orientation and construction site of the model only affects the results of exceedance probability. It should be noted that the results of the ACR correspond to the wind speed of 3 m/s (9.8 ft/s) at the building height.

[FIGURE 6 OMITTED]

[TABLE 1 OMITTED]

Figure 6 shows the experimental results of the ACR as a function of wind incident angle with different opening configurations. The result of the default configuration is presented in bold solid lines, while the configurations with two openings are shown in dashed lines. As can be seen from these curves, there are large variations of the ACR in individual cases at different wind incident angles. This indicates that natural ventilation performance in the basement is significantly affected by the wind direction, which should be taken into account in the analysis of natural ventilation potential. The curves of the configurations with one opening show the same behavior, i.e., high ACRs occur in the range of wind incidence angle from 90[degrees] to 135[degrees], while low values are found from 157.5[degrees] to 180[degrees]. A small difference in the incidence angle between 135[degrees] and 157.5[degrees] might result in a large difference in ventilation performance. By contrast, for the configurations with two openings, peak values are found at the incident angle of 90[degrees] when the wind is parallel to the opening facade. It can also be found from the same figure that the effect of increasing the opening size is obvious, and the ACRs of the configurations with two openings are apparently higher than with only one opening with the same opening size. The influence of opening configuration on natural ventilation potential in the basement will be discussed later based on the exceedance probability analysis.

The ACR results due to the changing of the areaway plan area are presented in Figure 7 as a function of wind incident angle. In all the four cases shown here, the opening configuration has one opening with an area of 3.24 [m.sup.2] (34.9 [ft.sup.2]), the areaway plan area is 18 [m.sup.2] (193.8 [ft.sup.2]), and the building coverage ratio is zero, i.e., the detached building pattern. Compared with the default configuration shown in solid lines, the curves of the other cases with smaller plan areas exhibit different behaviors at different wind incident angles. Generally, larger areaway plan area results in higher ACR. However, that is not true for certain wind directions. For instance, when [theta] = 90[degrees], the ACR is higher in the case for a narrower areaway space of A = 5.4 [m.sup.2] (58.1 [ft.sup.2])than for a wider areaway of A = 12.6 [m.sup.2] (135.6 [ft.sup.2]), which is probably related to the change of local flow pattern in the areaway space.

[FIGURE 7 OMITTED]

Figure 8 shows the results of the ACR as a function of wind incident angle with different building coverage ratios. The default configuration of R = 0 is presented in a solid line. It can be seen that compared with the detached building pattern (R = 0), the results of the multi-building pattern (R > 0) display different distributions as a function of wind direction. For R = 25% (in dotted-dashed line), the decreases in ACR can be found in the range of incident angles from 90[degrees] to 135[degrees], while a slight increase occurs from 22.5[degrees] to 45[degrees].For the other three multi-building cases, the influence of wind direction is relatively small.

Exceedance probability distribution and a simple analysis approach

Substituting the ACRs into Equation 4 and then into Equation 3 with local meteorological statistical data, one can get a total exceedance probability corresponding to a specific ACR value. Likewise, given a series of discrete values of ACR in a specific range, a plot of total exceedance probability distribution can be obtained.

[FIGURE 8 OMITTED]

[FIGURE 9 OMITTED]

As an illustration, the exceedance probability distributions plotted as a function of [ACR.sub.g,s] from 0 [h.sup.-1] to 40 [h.sup.-1] are presented in Figure 9 for the basement model with the default configuration and located in Tokyo. The calculations were performed for the basement model with 16 different orientations. For easy readability, Figure 9 only shows the distributions for two orientations with the overall highest and lowest probability values, i.e., E and SSE. The results for the other orientations generally lie between the two distribution curves. As indicated from Figure 9, the differences of exceedance probability between different orientations are small when [ACR.sub.g,s] is at high and low range of ACR, e.g. [ACR.sub.g,s] >35 [h.sup.-1] and [ACR.sub.g,s] < 5 [h.sup.-1]. However, such differences are considerable for [ACR.sub.g,s] in the middle range, e.g., it reaches 22% between orientation E and SSE at an ACR of 16 [h.sup.-1]. This is mainly attributed to the large variations of the ACR in the basement for each wind direction (see Figure 6) as well as the differences in the occurrence frequencies of wind direction in each azimuth sector (see Figure 5). For example, as far as a basement model with the orientation of east is concerned, as mentioned earlier, the wind incidence angles with high ACRs are in the range of 90[degrees] to 135[degrees], which cover the prevailing wind directions--NNW, N, and SW--and thus lead to the highest exceedance probability distribution. Similarly, when the orientation of the basement model is SSE, the wind incidence angles with low ACRs, 67.5[degrees], 157.5[degrees], and 180[degrees],are coincident with the three prevailing wind directions, resulting in the lowest exceedance probability distribution.

Although the exceedance probability distributions in Figure 9 provide a large amount of information concerning the changing tendency of natural ventilation potential, it is difficult to use these distribution curves to compare natural ventilation potential between different configurations directly and not convenient for practical assessment either. By contrast, it would be more straightforward to use the ACR index, which corresponds to some given exceedance probability values, for the quantitative analysis of natural ventilation potential. Assessment can further be made based on the comparison between these index values and specified thresholds, although this is beyond the scope of the present article. Here, a relative simple analysis approach is proposed for the analysis of natural ventilation potential by using the ACRs corresponding to exceedance probabilities of 1/7 and 6/7. From a realistic view, the exceedance probabilities of 1/7 and 6/7 are the probabilities of occurrence of one day per week and six days per week, based on the weekly activity pattern (Kato 2010). The ACR occurring one day per week indicates a high requirement for natural ventilation performance, while six days per week indicates a low requirement. See Kato (2010) for more detailed elaboration about this.

The detailed analysis approach is illustrated in Figure 10 for the basement model with the default configuration, where the values of ACR corresponding to exceedance probabilities of 15% ([approximately equal to]1/7) and 85% ([approximately equal to]6/7) are found from the respective distribution curves. As the ACR also changes with different model orientations, in order to ensure the worse case, only the lowest values are picked out from the corresponding distribution curves, hereinafter referred to as [ACR.sub.85%] and [ACR.sub.15%]. As depicted in Figure 10, [ACR.sub.15%] and [ACR.sub.85%] are 17.3 [h.sup.-1] and 5.3 [h.sup.-1] for the default basement model with NNE and south orientations, respectively, which means that the ACR of 17.3 [h.sup.-1] can be expected for about one day in a week in the areaway-attached basement with the worst orientation and 5.3 [h.sup.-1] for six days in a week. It should be noted that the model orientations of [ACR.sub.15%] and [ACR.sub.85%] might differ from the orientations with the overall lowest probability values.

Discussion based on exceedance probability analysis

Based on the exceedance probability distribution and the proposed simple analysis approach using [ACR.sub.15%] and [ACR.sub.85%], the parameters of opening configuration, areaway plan area, building coverage ratio, and site of the model were investigated in detail in this section.

[FIGURE 10 OMITTED]

The exceedance probability distributions of five opening configurations (Table 1) were depicted in Figures 11 and 12 for the models with orientations E and SSE, which correspond to the orientations with the overall highest and lowest exceedance probability values, respectively. Similar to Figure 9, the curves of the default configuration are presented in bold solid lines, while the configurations with two openings are in dashed lines. In both figures, the curves show similar trends as a function of ACR. It can also be seen that the curves of the two-opening configuration are always above those of one opening configurations with the same size, which demonstrates that the natural ventilation potential depends not only on the size of the opening but also the number of openings. [ACR.sub.15%] and [ACR.sub.85%] were further picked out from the respective probability distribution curves in Figure 12. Figure 13 shows these extracted values, which are arranged as a function of opening size. The approximate lines (dotted-dashed lines with zero j-intercept) are also presented according to the number of the opening. It can be seen that [ACR.sub.15%] and [ACR.sub.85%] are generally proportional to the opening size for the opening configurations with both one and two openings. For a basement with one smallest opening, an ACR of 2.5 [h.sup.-1] can be maintained during 85% of the time, at least, in the basement. In addition, increasing the number of the openings can effectively improve the natural ventilation performance. This is probably related to the separation distance between the two openings, which help increase the instantaneous pressure difference to drive the airflow through the openings. Similar results can also be found in Figure 11 in the literature (Bu et al. 2010), except that the previous results are mean ventilation rates averaged over nine wind directions, while the results of the present article are ACRs based on the exceedance probability analysis.

[FIGURE 11 OMITTED]

[FIGURE 12 OMITTED]

The influence of areaway plan area on the natural ventilation performance in the basement was further analyzed, and the results of exceedance probability distributions are presented in Figures 14 and 15 for the model orientations with the overall highest and lowest exceedance probability values in each case, respectively. It should be noted that although not shown in the figures, the orientations are different with areaway plan area. Compared with the distributions of the default configuration presented in bold solid lines, the other curves generally show similar trends as a function of ACR. However, for the configuration with the smallest plan area, the shape of the distribution is different and, it crosses another curve when [ACR.sub.g,s] [approximately equal to] 15 [h.sup.-1] (Figure 14), which probably results from the presence of a peak value of ACR at [theta] = 90[degrees] shown in Figure 7. The corresponding [ACR.sub.15%] and [ACR.sub.85%] extracted from Figure 15 are presented in Figure 16 along with two approximate lines (dotted-dashed lines with nonzero y-intercept) as a function of areaway plan area. It can be seen that fairly good linear relationships are also obtained between the ACR and the areaway plan area. Nevertheless, it should be noted that such linear relationships will not exist for a very small areaway. Compared with the default configuration with the largest areaway plan area investigated in the present article, [ACR.sub.15%] and [ACR.sub.85%] of A = 5.4 [m.sup.2] (58.1 [ft.sup.2]) decrease by 49% and 43%, respectively.

[FIGURE 13 OMITTED]

[FIGURE 14 OMITTED]

[FIGURE 15 OMITTED]

[FIGURE 16 OMITTED]

As presented in Figures 17 and 18, the exceedance probability distributions were also used to investigate the influence of building coverage ratio on natural ventilation potential in the areaway-attached basement. As in Figures 14 and 15, the results are also classified according to the model orientations with the overall highest and lowest exceedance probability values. It can be seen that in either figure, the exceedance probability values of the multi-building pattern (R > 0) are generally lower than the detached building pattern (R = 0). In addition, for each multi-building pattern, the differences are small in terms of probability values between the orientation with the largest exceedance probabilities and the orientation with the lowest ones. This implies that due to the presence of surrounding buildings in the multi-building patterns, the influence of the model orientation on natural ventilation potential becomes not as significant as in the detached building pattern. Figure 19 shows the extracted [ACR.sub.15%] and [ACR.sub.85%] arranged as a function of building coverage ratio, together with the corresponding approximate lines. Generally speaking, the increase in building coverage ratio results in a decrease in ACR to some extent. Compared with the detached building pattern, [ACR.sub.15%] and [ACR.sub.85%] of R = 33% are found to be 36% and 33% lower. However, it should be noted that in some cases, the ACR is higher with an increased coverage ratio (e.g., comparing R = 25% and 22%). According to Table 1, the only difference between the two patterns is in the reversed separation distances in the x and j directions. This implies that local air change phenomenon is quite complicated within a multi-building pattern, and more detailed airflow analysis advanced analysis using, e.g., CFD, is needed in future work.

[FIGURE 17 OMITTED]

[FIGURE 18 OMITTED]

Figure 20 shows a comparison of the exceedance probability distributions between three major cities in Japan: Tokyo, Osaka, and Sendai. In this figure, the results for the model orientations with both the overall highest and lowest exceedance probability values are presented together, marked in the legend as "Max" and "Min," respectively. It can be seen clearly that compared with Tokyo, the differences between the highest and lowest probability curves of Osaka and Sendai become smaller, especially in the range of [ACR.sub.g,s] from 0 [h.sup.-1] to 20 [h.sup.-1]. For instance, the maximum difference of exceedance probability value in Sendai is just 12% between orientation NE and NNW at an ACR of 20 [h.sup.-1]. This is mainly attributed to the differences in the wind rose of the three cities (Figure 5), as well as the differences of the Weibull parameters K and C (not shown here). As mentioned earlier, the discrepancies in the occurrence frequency of wind direction in each azimuth sector are more pronounced in Tokyo than in other cities. Therefore, the exceedance probability distribution is more sensitive to model orientation in Tokyo. Figure 21 shows the values of [ACR.sub.15%] and [ACR.sub.85%] in the three cities for the same basement model with the default configuration. Compared with Tokyo, [ACR.sub.85%] in Sendai decreases by 17% while [ACR.sub.15%] increases by 19%, and the values in Osaka is between those of Tokyo and Sendai. This demonstrates the influence of the construction site of the basement model on natural ventilation performance due to different wind conditions in these cities.

[FIGURE 19 OMITTED]

[FIGURE 20 OMITTED]

[FIGURE 21 OMITTED]

Conclusions

With the attention focused on single-sided natural ventilation driven by wind force, the concept of exceedance probability analysis was applied in the present article to investigate the natural ventilation potential in an areaway-attached basement. Different from the previously proposed criteria for assessing problems caused by strong wind at pedestrian level, the ACR in the target domain, i.e., in the areaway-attached basement space, was used as the evaluation index instead of wind velocity in the calculation of exceedance probability. This ACR-based exceedance probability can enable an evaluation of reliability in terms of natural ventilation potential and can take the uncertainties associated with stochastic wind in urban areas into full account. Based on the exceedance probability distribution curves, a simple analysis approach was proposed by extracting ACRs corresponding to the exceedance probabilities of 85% and 15% from the exceedance probability distribution curves, which is more convenient in practice for evaluating natural ventilation performance.

Based on the exceedance probability distribution and the proposed simple analysis approach, a variety of parameters regarding areaway-attached basement design were investigated in detail. The opening configuration and the areaway plan area show significant influence on ventilation performance in the basement. Good linear relationships were obtained between both parameters and the ACR. Moreover, increasing the number of the openings appears to be an effective means to improve the natural ventilation performance for single-sided ventilation. High building coverage ratio generally results in a decreased ACR, and the presence of surrounding buildings leads to a decrease in the difference of natural ventilation potential between different model orientations. However, the influence of building coverage ratio seems to be very complicated, which requires detailed airflow analysis using CFD in future work. In addition, natural ventilation potential in the basement is also influenced by construction site due to different urban wind conditions.

Although the areaway-attached basement model used in the present article is quite simple, the findings of this study can provide some useful information and guidelines for areaway design and application in practice. It should also be noted that thermal effects and human behavior are not included in the analysis of natural ventilation potential, which deserves further consideration.

DOI: 10.1080/10789669.2011.564261

Acknowledgments

This study is partially supported by a research grant from the Japan Society for the Promotion of Science (Project no. 09J09505) and partially by the Obayashi Foundation of Japan.

References

Al-Mumin, A.A. 2001. Suitability of sunken courtyards in the desert climate of Kuwait. Energy and Buildings 33(2):103-11.

Bu, Z., S. Kato, Y. Ishida, and H. Huang. 2009. New criteria for assessing local wind environment at pedestrian level based on exceedance probability analysis. Building and Environment 44(7):1501-8.

Bu, Z., S. Kato, and T. Takahashi. 2010. Wind tunnel experiments on wind-induced natural ventilation rate in residential basements with areaway space. Building and Environment 45(10):2263-72.

Hall, D.J., S. Walker, and A.M. Spanton. 1999. Dispersion from courtyards and other enclosed spaces. Atmospheric Environment 33(8):1187-203.

Hunt, J.C.R., E.C. Poulton, and J.C. Mumford. 1976. The effects of wind on people; new criteria based on wind tunnel experiments. Building and Environment 11(1): 15-28.

Kato S. 2010. urban wind environment and its impact on indoor environment - acceptable wind features of void space surrounded by buildings. Proceedings of the 7th International Conference on Indoor Air Quality Ventilation and Energy Conservation in Buildings, August 16, 2010, Syracuse, New York, pp. 193-448.

Kato S., K. Ito, and S. Murakami. 2003. Analysis of visitation frequency through particle tracking method based on LES and model experiment. Indoor Air 13(2):182-93.

Larsen, T.S., and P. Heiselberg. 2008. Single-sided natural ventilation driven by wind pressure and temperature difference. Energy and Buildings 40(6):1031-40.

Melbourne, W.H. 1978. Criteria for environmental wind conditions. Journal of Wind Engineering and Industrial Aerodynamics 3:241-9.

Murakami, S., Y. Iwasa, and Y. Morikawa. 1986. Study on acceptable criteria for assessing wind environment at ground level based on residents' diaries. Journal of Wind Engineering and Industrial Aerodynamics 24(1):1-18.

Narasaki, M., T. Yamanaka, and M. Higuchi. 1989. Influence of turbulent wind on the ventilation of an enclosure with a single opening. Environment International 15(1-6): 627-34.

Pietrzyk, K., and C.-E. Hagentoft. 2008. Probabilistic analysis of air infiltration in low-rise buildings. Building and Environment 43(4):537-49.

Rajapaksha, I., H. Nagai, and M. Okumiya. 2003. A ventilated courtyard as a passive cooling strategy in the warm humid tropics. RenewableEnergy28(11):1755-78.

Ratcliff, M.A., and J.A. Peterka. 1990. Comparison of pedestrian wind acceptability criteria. Journal of Wind Engineering and Industrial Aerodynamics 36: 791-800.

Russell, M., M. Sherman, and A. Rudd. 2007. Review of residential ventilation technologies. HVAC&R Research 13(2):325-48.

Warren, P.R., and L.M. Parkins. 1985. Single-sided ventilation through open windows. Proceedings Thermal Performance of the Exterior Envelopes of Buildings Florida ASHRAESP 49:209-28.

Zhen Bu (1,2) * and Shinsuke Kato (1)

(1) Institute of Industrial Science, University of Tokyo, CW405, 4-6-1, Komaba, Meguro-Ku, Tokyo 153-8505, Japan

(2) Mott Macdonald Ltd, Shanghai 200030, China

* Corresponding author e-mail: buzhen@iis.u-tokyo.ac.jp

Received December 1, 2010; accepted February 4, 2011

Zhen Bu, PhD, is Post-doctoral Researcher and Senior Building Physics Specialist. Shinsuke Kato, PhD, Fellow Member ASHRAE, is Professor.

Printer friendly Cite/link Email Feedback | |

Author: | Bu, Zhen; Kato, Shinsuke |
---|---|

Publication: | HVAC & R Research |

Article Type: | Report |

Geographic Code: | 9JAPA |

Date: | Sep 1, 2011 |

Words: | 5875 |

Previous Article: | Thermal comfort and indoor air quality in rooms with integrated personalized ventilation and under-floor air distribution systems. |

Next Article: | Whole-building heat, air, and moisture transfer modeling for residential buildings in different climates. |

Topics: |