An analysis of wind conditions at pedestrian level in the selected types of multi-family housing developments

The following article addresses the issue of wind conditions around urban building development at pedestrian level. Factors that depend on those issues include wind comfort and air quality within urbanized spaces. The conditions specific of cities located in a temperate climate zone have been taken into account. The article is intended to identify aerodynamic phenomena characteristic of the three basic types of multi-family building development (section, linear, quarter). For each type of buildings, three different variants were considered on the basis of land use efficiency (low, medium, high) each type of buildings is characterised with. The models illustrate the patterns observed within European urban planning thought and have been developed in accordance with contemporary architectural and urban conditions, as well as with legal requirements and good practices in force in Poland. Experimental research in the wind tunnel has been conducted by means of two complementary field methods (visualization and erosion). The application of these two methods made it possible to obtain data on airflow direction layouts and on changes in airflow velocity. Subsequently, the results obtained in the study have been compared. Thus, more and less favourable geometrical layouts in respect to aerodynamic phenomena have been indicated.


Introduction
Development of modern big cities gives rise to a number of issues, one of which seems especially serious, namely the need to improve the quality of big city mesoclimate conditions. Mesoclimate conditions are related to wind phenomena, which play a vital role in ensuring proper ventilation of urban spaces, allow for temperature regulation and pollution removal. In temperate climate zone, a necessity arises to limit the occurrence of phenomena that either excessively intensify airflow velocity (for the sake of pedestrian comfort and for protection of buildings against excessive cooling) or that limit air circulation. The latter goal is crucial in order to avoid the problem of urban heat island, as well as to minimise the problem of smog, both of which are troublesome for big cities.
Research on airflow around buildings has been intensively developing since the midtwentieth century [3,4,19]. As to a particular research that accounted for a breakthrough in the application of airflow research, J. Gandemer's experimental research from the 1970s should to be mentioned. The research made it possible to distinguish twelve types of aerodynamic effects that occur around isolated buildings and around simple layouts of such buildings [14]. In the course of later studies, these phenomena were recognized more accurately by other authors [19]. Studies on street canyons conducted in the 1990s were especially important for urban planning. For example R. Walker, L. Shao and M. Wolliscroft conducted one of the first studies of natural ventilation via courtyards based on computational fluid dynamics [34]. These studies offered a better understanding of the relationship between airflow in urban spaces and the geometric form of complexes of buildings. Baik and Kim explored air circulation and the emergence of multiple vortices focusing on skimming flow relations in narrow courtyards [1]. The dependence of air circulation on the W/H parameter was observed and approximated figures for this correlation were provided, wherein the W/H parameter was expressed in terms of the distance between buildings (W) and their height (H). This knowledge has been further developed by many other authors and used as the basis for principles and recommendations for pro-ecological design consistent with the idea of sustainable development [11].
Research possibilities concerning aerodynamic phenomena around buildings have expanded considerably over the last 20 years, along with the development of numerical simulation methods. The principles for the application of these methods, as well as principles applied in order to combine the methods with experiments have been widely described in source literature [2,15]. Most commonly, objects with simple geometrical layouts: free standing [20,21,23,29] or regular layout of solids [3,27] provide the research topic. Publications concerning more complex geometrical layouts that take the surroundings of a building into account are still relatively rare [5,19,30,31,35]. Such publications are dominated by case studies and concern specific types of building developments. Attempts to discuss model solutions that reflect the complex conditions commonly experienced in the process of urban development design are less numerous [18,28,31]. However, such publications seem particularly valid from the point of view of urban planning. Perhaps the lack of sufficient amount of publications on the topic is one of the reasons behind an infrequent application of the wind engineering theory in the practice of urban and architectural design.
The aforementioned previous studies clearly indicate that the airflow around buildings is closely related to the shape of a particular building development-the sizes and forms of individual buildings and their positioning in relation to one another. Several factors play a vital role for ventilation. They include: geometric proportions of streets [1,8,11,34,36], variations in building height [9,26], building intensity expressed in Floor Aspect Ratio and Building Coverage Ratio [25,29]. Moreover, geometric features, though not sufficiently tested yet, may exert an impact on the efficiency of urban ventilation [19]. These features include: porosity (the amount of empty spaces of an urban layout in which airflow can occur), orientation of the network of streets (related to the influence of building and street orientation on airflows), share of sheltered spaces (enclosures as a courtyards or 1 3 narrow, closed streets) [19]. Therefore, designing the shape of a new building is related to a modification of the aerodynamic phenomena in its surroundings. The conscious use of this dependence in urban and architectural design can positively affect the quality of the microclimate in spaces between buildings (in squares, streets, courtyards, parks, gardens, etc.), as well as it may indirectly influence the overall mesoclimate situation of the city.
The following article is intended to analyse aerodynamic phenomena that occur at pedestrian level around the three most common types of multi-family housing developments. Differences between individual variants were observed, whereas wind conditions that constitute pedestrian comfort within spaces between buildings and that account for the possibility of airing spaces between the buildings were compared. The models adopted for the present research exemplify the types of buildings most commonly seen in urbanized areas in a temperate climate zone. The authors' objective was to create such model objects that would be close to real issues observable in architectural and urban design, especially in case of Polish conditions (although the models can probably be applied to urbanized spaces of many other countries).

The subject of the research
Three types of multi-family building developments have been selected as the subject of the present research: • Section developments -free-standing buildings with rectangular, but not overly elongated building plans, • Linear developments -buildings with a considerably elongated rectangular plan, • Quarter developments -a set of buildings that enclose a rectangular inner courtyard.
The aforementioned types of multi-family building developments are characteristic of the European architectural thought fostered since the beginning of the 20th century and continued to the present day. Such building developments can be found in the urban tissue of many Polish cities ( Fig. 1) [16,24].
Pallado, an architect, established theoretical geometric models that correspond to the above types of building developments [24]. The models illustrate land development of a square plot, covering an area of 2 hectares, on which an adequate amount of recreational and green spaces has been planned with reference to inhabitant numbers, as well as a proper amount of spaces for communication and parking purposes has been intended. Legal conditions in force in Poland have been taken into account (requirements for room insolation and lighting with daylight, for dimensions and position of roads, walkways, parking lots, etc.) and good design practices have been applied. Each model has been developed in three variants -for low, medium and high degree of land use efficiency (LUE). The land use efficiency has been defined by the three parameters enumerated below (the range of values has been provided in Table 1): • Area use expressed in the number of flats per hectare of plot area, • Population density expressed in the number of residents per hectare of plot area, • Building development intensity expressed as the ratio of the total area of building developments (all above-ground storeys) to the area of the plot on which the development is situated.
The buildings in the models developed by Pallado (and applied in this research) are of the following heights ( Fig. 2): 1 3 • Section building development -12 m (4 storeys) for low LUE, 33 m (11 storeys) for medium LUE, 27 m (9 storeys) for high LUE, • Linear building development -9 m (3 storeys) for low LUE, 24 m (8 storeys) for medium LUE, 27 m (9 storeys) for high LUE, • Quarter building development -12 m (4 storeys) for low LUE, 21 m (7 storeys) for medium LUE, 18 m (6 storeys) for high LUE; in this case gateways were introduced on two opposite sides of the block (4 for low and medium LUE and 6 for high LUE).
Additionally, the research presented in the article takes into account the surrounding building development so as to recreate the urban context of exemplary layouts of building developments. In order to recreate the surrounding building development, elongated buildings of 9 m (3 storeys) in height have been chosen to form four streets adjacent to a square plot. A uniform environment has been assumed for all of the layouts of building developments tested.

Research techniques
The research was conducted in an aerodynamic tunnel whose measuring space equals 1 m × 1 m. In order to depict properly the ground floor level, a ready-made set of obstacles has been introduced. The obstacles had been specifically designed in accordance with wind standards applied in the studies described in [6,10,12,15,17] and elsewhere. In this way, the desired profile of velocity, intensity and kinetic energy of turbulence were obtained. The building models adopted for research were created in 1:400 scale.
Two field techniques, namely visualization method and erosion method, have been applied to observe the aerodynamic phenomena that arise in the discussed urban layouts and to determine changes in wind velocity in immediate vicinity of the buildings. The two techniques complement each other. The first one determines the flow direction, the second one determines the changes in flow velocity. The advantage of the two methods may be seen in their simplicity and in relatively low performance costs. The results yielded by the two methods, though of qualitative nature, do accurately reflect the flow image and allow for a meticulous analysis [3,6]. The areas where airflow accelerates, as well as areas of poor ventilation are clearly visible in the results. Moreover, it is possible to estimate the value of alteration in velocity in specific areas.
Of all visualization techniques available, the oil visualisation technique was chosen. In order to conduct research described in the article, a glass base, in black, to which the models were attached has been used. The glass base was painted with a mixture of oil and titanium dioxide, which served as a white marker (Fig. 3).
As a result of simulation in the wind tunnel, an image of the averaged moving air directions, as well as of a layout of vortex structures in the flow have been obtained. During the experiment, photos were taken at a frequency of four photos per minute. The analysis of the sequence of all the photos taken in a given trial allowed for a full understanding of the phenomena that occur in the flow and for a correct interpretation of such phenomena. The interpretation of the obtained phenomena was marked with graphical symbols in the last picture in each sequence (Figs. 6 and 8).

3
The visualization technique was further supplemented with the erosion techniquenamely, with sand erosion. The application of this method makes it possible to determine changes in airflow velocity values once an obstacle has been introduced (in the described studies such an obstacle is posed by layouts of buildings). A detailed description of the abovementioned technique can be found in [3,6].
The experiment in which the erosion technique is applied consists of two stages. In the first stage, the reference speed should be determined. In this case, the plate with no models on it is evenly sprinkled with sand, while airflow velocity at which the layer of applied sand is almost completely blown off the plate is measured in the tunnel. Then, a plate with models of buildings attached to it and with sand spread evenly on the empty areas between the buildings is placed in the tunnel. Once the tunnel is launched, airflow velocity in it is gradually increased. After each alteration of speed, the airflow stabilizes, then a photo of the examined area is taken. In accordance with the principles behind the use of this technique, twelve pictures were applied and processed for each variant (for subsequent airflow velocities (Fig. 4). As a result, coloured maps that display distribution of wind amplification coefficient marked as α have been obtained (Fig. 5).
The relation defines this coefficient, where U GB denotes the ground-level wind speed that is influenced by the building and for which sand erosion occurs, and U G,n denotes the ground-level wind speed that is not influenced by the building. (Ground-level wind speed corresponds to the speed measured at the pedestrian level, i.e. at a height of about 1.75 m).
In the wind tunnel, all dimensions are rendered in a scale. For a tunnel that measures 1 m x 1 m, a height of 1.75 m corresponds to a height of 4.4 mm. It is difficult to measure the velocity accurately at this height, it is much easier to measure tunnel operating speed. It can be shown that to determine wind amplification coefficient it is enough to measure the speed anywhere in the tunnel.
Let U W stand for the wind velocity measured at a certain height from the base of the tunnel (in our case this height was 0, 6 m). The velocity is gradually increased until the process of erosion begins and almost all the sand is blown off the plate. The value of U WB   07-1.15)-zones of lower intensity of ventilation, at average wind force, they create comfortable conditions for residents, whereas at very strong winds they are no longer as troublesome as red zones; α∈ (0.83-1)-zones of good ventilation, both at low and high wind velocity; α∈ (0.71-0.77)-zones of reduced airflow, which at regular wind conditions are characterized by low ventilation, relatively good at strong winds; α∈ (0.66-0.00) zones of poor recirculation and a stagnation zone is then proportional to the velocity at the ground-level, U GB , at which the wind actually affects the sand. In the following stage of the research, buildings are attached to the cleaned plate and a layer of sand is applied once again. The flow velocity is increased gradually to the values of U W,1 , U W,2 , U W,3 , … Before each modification to wind velocity in the tunnel, some amount of time is devoted to determine the shape of the area of the blown-off sand. These velocities correspond to velocities observable at ground-level U G,1 , U G,2 , U G,3 … An acceleration of flow near buildings may occur. In areas where the airflow velocity exceeds the U GB value, sand is blown off. It should be noted that using dependencies The wind amplification coefficient may also be written in the form of: It should be noted that in the area where > 1, airflow accelerates, e.g., due to the presence of buildings. Conversely, for < 1 airflow slows down.
When the average or maximum wind velocity (V) in the city is identified at the pedestrian level, an approximate velocity in the presented built-up areas can be determined as w = αv, based on the wind amplification coefficient.
Wind velocity in the climatic conditions in Poland fluctuates on an annual scale from 1.16 to 12.09 m/s, the average velocity equals about 3.5 m/s, while in case of a city, such as Warsaw, it reaches about 4 m/s [13]. Taking into account the wind comfort conditions presented in numerous publications [7,8,17,22,32,33], the colour layout may be interpreted as described in the Fig. 5.

Method of research grouping and result comparison
Each of the nine building development variants has been tested with the application of two techniques described above. Additionally, for each of them, three wind directionsparalleled, perpendicular and oblique (at 45° angle) were taken into account. Due to the almost complete symmetry of the building development layouts, it may be concluded that the results for winds of the same directions, but opposite turns would provide a symmetrical reflection.
Altogether, twenty-seven tests have been performed with the application of both techniques, the results have been compared using the following steps: a. The results for one type of building development have been analysed (starting with the section building development) for low, medium and high land use efficiency, in case of one (paralleled) wind direction. The results have then been compared; b. Analogous analyses and comparisons have been conducted, as described above, for the remaining wind directions (perpendicular and oblique ones); c. The results have been compared and a collective comparison has been made of the variants for one building development type (starting with the section development-three variants of land use efficiency and three wind directions); d. analogous actions have been performed (as in points a, b and c) with respect to the results obtained for other types of building developments (linear and quarter ones); e. The synthesized results for three types of building developments have been juxtaposed and compared in three variants of land use efficiency.

Results
Owing to a large number of partial results obtained in the study, an exemplary interpretation of the results for one type of building development, namely the quarter building development with a high land use efficiency in case of paralleled wind direction has been presented in the article (see 4.1.1). Moreover, results for all three land use efficiency variants in case of the abovementioned wind direction have been holistically compared (see 4.1.2). Subsequently, a summary of results for this type of building development has been described, using three variants of land use efficiency for three wind directions (see 4

.1.3).
The results for other types of buildings, that is, section (see 4.2) and linear (see 4.3) have been described in a synthetized manner, comparing all three variants of land use efficiency for the three wind directions analysed in the study (the description of detailed comparisons of research results has been omitted here). Following that stage, a final comparison of all variants has been presented (see 4.4).

Results for quarter building development with high land use efficiencyparalleled wind direction
Based on the results of visualization study (Fig. 6) and erosion study (Fig. 7), it can be observed that the main airflow reaches the interior of the layout between the O1 and O2 buildings. The airflow bounces off the PW1 wall of the main building, which results in the incoming air being slowed down. Next, the main air flow splits into two accelerating streams flowing along the O1 and O2 buildings and falling into the spaces between the main PW building and the O3 and O4 buildings. A large stagnation zone becomes noticeable in the courtyard of the main building, although streams of air that flow in through the gates of the courtyard can be seen. The air is rebounded off the windward side of the PW2 wing, and then moves, gradually slowing in velocity, to the central part of the courtyard. Clear stagnation zones can be observed on the leeward side of the PW building and the O3, O4, O5, O6 buildings. The zones of highest flow velocity appear in the corners of the windward edges of the O1 and O2 buildings and at the windward walls of the main building (PW1), as well as in the spaces between the buildings mentioned. Even when wind velocity reaches values lower than 5 m/s, these zones may prove uncomfortable for pedestrians, as the velocity there is intensified by 25 to 36%.

Comparison of results for quarter building development according to three variants of land use efficiency-paralleled wind direction
Likewise, the airflow for quarter building development with low and medium land use efficiency has been analysed. Following the analysis, the results have been juxtaposed for comparison. When analysing the consolidated summary of the study (Fig. 8), some common features and differences can be observed. The largest stagnation zone inside a courtyard is formed in the case of building developments with medium land use efficiency (due to the fact that the courtyard in this case is the "tightest", the coefficient equals W/H = 2.57). Whereas the weakest stagnation zone occurs in the case of developments with high land use efficiency (W/H ratio of the courtyard = 4.89). Outside the inner area of the courtyard, the largest zone of a low amplification coefficient (α∈ (0.0.66)) forms also for buildings with medium land use efficiency, while the weakest stagnation zone is visible in the case of building developments with high land use efficiency. The largest areas of intensified flow (α > 1.25) occur in vicinity of the windward side of the main building for building developments of high land use efficiency.

Comparison of results for quarter building development in three variants of land use efficiency-paralleled, perpendicular and oblique wind direction
Similar research has been conducted for perpendicular and the oblique wind directions. Quarter building development with low land use efficiency is characterized in all cases by low level of airiness for the building quarter and for its surroundings. The courtyard of a quarter building development with a medium land use efficiency is clearly less ventilated, while the space between neighbouring buildings is better ventilated.
In the case of the quarter building development with a high land use efficiency, the courtyard is relatively large and best ventilated, as opposed to the previously discussed variants. Yet, in the space between neighbouring buildings, there exist areas with a higher rate of velocity gain (α > 1.25). These areas, however, are not extensive. For areas affected by winds of velocities higher than 4 m/s, a variant with an medium land use efficiency seems more comfortable to pedestrians. Still, if wind velocities in a given area are rarely on the increase, quarter building development with high land use efficiency seems the most favourable one, as it offers the best airiness of the entire quarter.

Comparison of results for section building developments in three variants of land use efficiency-paralleled, perpendicular and oblique wind direction
Analogous tests and comparisons have been performed for section building developments with low, medium and high land use efficiency. Section building development with low . Whenever a stronger wind is blowing (above 4 m/s), section building development of high land use efficiency turns into a less comfortable layout. This happens due to the occurrence of larger areas with a higher wind amplification factor (α > 1.25).

Final results of all the variants comparison
Research conducted with the use of oil visualization and sand erosion methods for all three types of building developments in three variants of land use efficiency have been compared. In all of the layouts, minor areas with an increased amplification factor (α > 1.25, which appear mainly at the corners of buildings) are observable, together with less and more extensive stagnation zones. In the building development models tested in the present research, the issue of limited ventilation possibilities poses greater problems than the issue of the wind acceleration that may prove uncomfortable for pedestrians. Research has shown that selection of the optimal building development layouts, considering various wind directions, is anything but easy. Having analysed the results obtained with the use of visualization and erosion techniques, more and less advantageous configurations have been selected (Table 2). Figures 9, 10, and 11 present visualizations of results obtained by erosion technique.
It has been determined that less favourable wind conditions in the pedestrian zone prevail around building developments of low land use efficiency in cases of section, linear and quarter building developments. Section building development seems the least favourable one. In this case, the largest areas of reduced velocity have been observed, for instance, ones for which α ≤ 0.66, mainly due to the presence of a number of buildings in the plot and small spaces between objects. These spaces are rather poorly ventilated and thus belong to potential stagnation zones, where pollution can accumulate. On the other hand, many local spaces where the velocity is amplified by α > 1.25 may be noticed. This is linked to the phenomenon of airflow acceleration at building corners. The more blocks of buildings are introduced in the plot, the more local zones of increased velocity appear. For section and linear building developments, more favourable configurations of medium land use efficiency have been discovered, where the majority of the area is well ventilated, with minor stagnation zones. For the quarter building development, it has been determined that the most favourable layout was the one of a high-efficiency building development, with the largest courtyard and with relatively best air exchange within it.
Having analysed all of the types of building developments, linear building development of medium efficiency seems to be the most favourable one when it comes to wind conditions in pedestrian zone. For all of the wind directions studied in the present research, the observed stagnation zones were rather insignificant. The geometric structure of the linear building development is more open to air inflow than the quarter building developments with closed courtyards. Low-efficiency section building development, on the other hand, has been identified as the least favourable of all types of the layouts tested, as it is the most fragmented one. It also contains the largest number of corners. The results obtained while testing this type of layout have pointed to large zones of poorly ventilated areas where pollution accumulates. An intensified flow was observed at all the edges of the buildings.

Discussion
The obtained results were of varied nature and were thus difficult to use for a clear comparative assessment. However, the conclusions derived from the comparison of individual variants of land use efficiency within one type of building development proved more unambiguous. It is worth noting that the obtained detailed results allow for comparisons between any groups of variants. Moreover, the results might prove useful, for example, to draw comparisons between different types of building developments for one variant of land use efficiency. Furthermore, it is valuable to analyse individual variants and see them as the ability to identify specific local phenomena.
It is important to reckon with the fact that when environment layout of a model building development is altered, the test results may differ, though probably not substantially so, assuming that the surrounding building development is compact and encircles the tested area. Problems were observed mainly regarding the layouts in which buildings are located close to each other. Increasing the height of the surrounding building development does not seem to essentially affect the elimination of these problems. On the contrary, such problems may be intensified due to obscuring. Formation of the surrounding building development as a less compact layout open to the inlet of airflow into the studied area might cause a greater discrepancy in test results.

Conclusions
The research discussed in the present paper has made it possible to compare wind conditions in the spaces between buildings for three types of housing development for low, medium and high land use efficiency. It should be emphasized that the results concern only theoretical models developed by Pallado [24]. The results indicate that, regardless of the type of building development, more favourable wind conditions at pedestrian level prevail around building developments of medium and high land use efficiency. This happens due to greater height of development layouts, as well as owing to the need to provide a larger amount of undeveloped areas in their vicinity (recreational areas, road layout, car parks).
The results, systematized in groups at various levels of detailedness, provide numerous conclusions that may be useful in making urban planning decisions regarding the use of areas for a specific type of building development and architectural decisions concerning building formation. It seems purposeful to develop this type of research in relation to other types of building layouts characteristic for heavily urbanized areas, as they tend to be affected to the greatest degree by problems in the field of air pollution and summer overheating. The method presented here is relevant in relation to model situations, the analysis of which provides general knowledge. The method is also useful as a support in the design process, but due to its time-consuming nature, it can be used to check selected variants of situations. A stronger integration of wind research with design process would require a proper combination of elements of the above described method with numerical techniques. 1 This approach would allow for faster action, better synchronized with the spatial decision-making process.
Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. 1 For several cases presented in here, numerical simulations with the use of the ANSYS FLUENT package have been conducted. The results obtained from such simulations proved to be comparable in terms of quality to the experimental research results and they lead to similar conclusions. .