Natural Ventilation E ﬀ ectiveness of Awning Windows in Restrooms in K-12 Public Schools

: Using computational ﬂuid dynamics (CFD), this study explores the e ﬀ ect of a di ﬀ erent number of awning windows and their installation locations on the airﬂow patterns and air contaminant distributions in restrooms in K-12 (for kindergarten to 12th grade) public schools in Taiwan. A representative restroom conﬁguration with dimensions of 10.65 m × 9.2 m × 3.2 m (height) was selected as the investigated object. Based on the façade design feasibility, seven possible awning window conﬁgurations were considered. The results indicate that an adequate number of windows and appropriate installation locations are required to ensure the natural ventilation e ﬀ ectiveness of awning windows. The recommended installation conﬁguration is provided.


Introduction
In Taiwan, restrooms in K-12 public schools (kindergarten (K) and the 1st through the 12th grade (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)) are frequently accessed and are open to the public, which hinders their management. In particular, restrooms that have been in service for more than 20 years often exhibit problems, such as poor ventilation, inadequate lighting, outdated layouts and designs and limited toilet spaces. Thus, these restrooms have become a nuisance for school environment management. Poor-quality restrooms create fear of using the restroom among teachers and students, which increases incidents in which teachers and students are affected by acute or chronic diseases that can adversely affect their physical and mental health.
There are few studies on bathroom and restroom ventilation that address the ventilation efficiency of entire bathrooms and restrooms as well as the ventilation efficiency of their components (e.g., toilets and fans). Investigating the ventilation of an entire bathroom and restroom, Tung et al. [1] analyzed a new negative pressure wall-exhaust ventilation system (that differs from the traditional ceiling-exhaust system) installed in a residential bathroom and restroom using a full-scale test. In the test, toilets were deployed in several different patterns and positions. The test results indicated that the restroom ventilation system could take advantage of a negative pressure difference to prevent the escape of restroom malodor to an adjacent room. Increasing the ventilation volume or decreasing the distance between vent and toilet could improve both the indoor pollutant removal rate and the ventilation rate. From the perspective of energy efficiency, an air change rate of 8.5 h −1 was the optimal value. Tung et al. [2] analyzed the effectiveness of three ventilation strategies for residential bathroom malodor removal: (1) forced ceiling-supply and wall-exhaust systems, (2) natural window-inlet and forced ceiling-exhaust systems and (3) forced ceiling-supply and ceiling-exhaust systems. The first strategy achieved the best malodor removal. Yang and Kim [3] employed computational fluid dynamics (CFD) to analyze the effect of changing the glass partition shape in an apartment bathroom on the bathroom   (Table 1), which is located on a certain floor and has seven window configurations (W1-W7). The windows for men's and women's restrooms are installed symmetrically. Whether a window frame is listed as a CFD simulation object affects the number of adjustable opening angle to control the ventilation area and deflect rain (which is particularly suitable for the typhoon and rainy season in Taiwan). Therefore, the awning window (as illustrated in Figure  2) is selected as the window configuration for the subject.   (Table 1), which is located on a certain floor and has seven window configurations (W1-W7). The windows for men's and women's restrooms are installed symmetrically. Whether a window frame is listed as a CFD simulation object affects the number of grids in the simulation and may significantly increase computing time. Therefore, the glazing dimensions are used as the dimensions of the ventilation opening; i.e., the window frame is ignored here. The installation of four awning windows comprises three designs (the black blocks in Table 1bd). The installation of eight awning windows consists of three designs (Table 1(e)-(g)). In addition, the installation of 12 awning windows has one design (Table 1(h)). The opening angles θ for the awning windows are 30°, 45° and 60°. The approaching wind speeds are 0.5 m/s, 1 m/s and 2 m/s. In  (Table 1), which is located on a certain floor and has seven window configurations (W1-W7). The windows for men's and women's restrooms are installed symmetrically. Whether a window frame is listed as a CFD simulation object affects the number of grids in the simulation and may significantly increase computing time. Therefore, the glazing dimensions are used as the dimensions of the ventilation opening; i.e., the window frame is ignored here. The installation of four awning windows comprises three designs (the black blocks in Table 1b-d).
The installation of eight awning windows consists of three designs (Table 1e-g). In addition, the installation of 12 awning windows has one design (Table 1h). The opening angles θ for the awning windows are 30 • , 45 • and 60 • . The approaching wind speeds are 0.5 m/s, 1 m/s and 2 m/s. In Taiwan, the weather and climate are greatly affected by monsoons. In summer, the prevailing wind is a southwesterly or southerly monsoon, while in winter it is northeasterly or northerly monsoon. Thus, the wind directions set in the CFD simulation include south and north winds. There is no heat source in the restroom, and the toilet doors are closed to simulate the scenario when the restroom is in use. Geometric data for the model under investigation are listed in Table 2. is a southwesterly or southerly monsoon, while in winter it is northeasterly or northerly monsoon. Thus, the wind directions set in the CFD simulation include south and north winds. There is no heat source in the restroom, and the toilet doors are closed to simulate the scenario when the restroom is in use. Geometric data for the model under investigation are listed in Table 2. Table 1. Quantity and position of installed awning windows (north facade).

Windows Configuration Legend Windows Configuration Legend
(a) North facade and opening angle.   The major cause of poor air quality in restrooms is the malodor of feces and urine. The smell of feces and urine is primarily the smell of ammonia. In contrast, sewer malodor consists of mixed chemical substances, such as hydrogen sulfide, methyl mercaptan, trimethylamine, dimethyl disulfide, indole and methyl indole. This study focuses on the restroom flow field and the malodor concentration field. Because the irritating odor of ammonia is believed to be a major contributor to the offensive odor of human waste [5], the NH3 concentration in the malodor is analyzed. To simulate the very worst condition, the generation rate of unpleasant odors (represented by NH3) was assumed to be 0.3 L/min (0.2 g (NH3)/min) in this study. The pollution source area is set to 0.1 m × 0.1 m.

Parts of the Model Geometric Data
A modified odor removal efficiency (ORE) [2,16,17] was employed to express the ventilation performance of the whole restroom. A higher ORE indicates a lower concentration level, thus indicating better ventilation efficiency for odor removal. The ORE is defined as: is a southwesterly or southerly monsoon, while in winter it is northeasterly or northerly monsoon. Thus, the wind directions set in the CFD simulation include south and north winds. There is no heat source in the restroom, and the toilet doors are closed to simulate the scenario when the restroom is in use. Geometric data for the model under investigation are listed in Table 2. Table 1. Quantity and position of installed awning windows (north facade).

Windows Configuration Legend Windows Configuration Legend
(a) North facade and opening angle.  The major cause of poor air quality in restrooms is the malodor of feces and urine. The smell of feces and urine is primarily the smell of ammonia. In contrast, sewer malodor consists of mixed chemical substances, such as hydrogen sulfide, methyl mercaptan, trimethylamine, dimethyl disulfide, indole and methyl indole. This study focuses on the restroom flow field and the malodor concentration field. Because the irritating odor of ammonia is believed to be a major contributor to the offensive odor of human waste [5], the NH3 concentration in the malodor is analyzed. To simulate the very worst condition, the generation rate of unpleasant odors (represented by NH3) was assumed to be 0.3 L/min (0.2 g (NH3)/min) in this study. The pollution source area is set to 0.1 m × 0.1 m.
A modified odor removal efficiency (ORE) [2,16,17] was employed to express the ventilation performance of the whole restroom. A higher ORE indicates a lower concentration level, thus indicating better ventilation efficiency for odor removal. The ORE is defined as: is a southwesterly or southerly monsoon, while in winter it is northeasterly or northerly monsoon. Thus, the wind directions set in the CFD simulation include south and north winds. There is no heat source in the restroom, and the toilet doors are closed to simulate the scenario when the restroom is in use. Geometric data for the model under investigation are listed in Table 2.  The major cause of poor air quality in restrooms is the malodor of feces and urine. The smell of feces and urine is primarily the smell of ammonia. In contrast, sewer malodor consists of mixed chemical substances, such as hydrogen sulfide, methyl mercaptan, trimethylamine, dimethyl disulfide, indole and methyl indole. This study focuses on the restroom flow field and the malodor concentration field. Because the irritating odor of ammonia is believed to be a major contributor to the offensive odor of human waste [5], the NH3 concentration in the malodor is analyzed. To simulate the very worst condition, the generation rate of unpleasant odors (represented by NH3) was assumed to be 0.3 L/min (0.2 g (NH3)/min) in this study. The pollution source area is set to 0.1 m × 0.1 m.
A modified odor removal efficiency (ORE) [2,16,17] was employed to express the ventilation performance of the whole restroom. A higher ORE indicates a lower concentration level, thus indicating better ventilation efficiency for odor removal. The ORE is defined as: is a southwesterly or southerly monsoon, while in winter it is northeasterly or northerly monsoon. Thus, the wind directions set in the CFD simulation include south and north winds. There is no heat source in the restroom, and the toilet doors are closed to simulate the scenario when the restroom is in use. Geometric data for the model under investigation are listed in Table 2.  The major cause of poor air quality in restrooms is the malodor of feces and urine. The smell of feces and urine is primarily the smell of ammonia. In contrast, sewer malodor consists of mixed chemical substances, such as hydrogen sulfide, methyl mercaptan, trimethylamine, dimethyl disulfide, indole and methyl indole. This study focuses on the restroom flow field and the malodor concentration field. Because the irritating odor of ammonia is believed to be a major contributor to the offensive odor of human waste [5], the NH3 concentration in the malodor is analyzed. To simulate the very worst condition, the generation rate of unpleasant odors (represented by NH3) was assumed to be 0.3 L/min (0.2 g (NH3)/min) in this study. The pollution source area is set to 0.1 m × 0.1 m.
A modified odor removal efficiency (ORE) [2,16,17] was employed to express the ventilation performance of the whole restroom. A higher ORE indicates a lower concentration level, thus indicating better ventilation efficiency for odor removal. The ORE is defined as: is a southwesterly or southerly monsoon, while in winter it is northeasterly or northerly monsoon. Thus, the wind directions set in the CFD simulation include south and north winds. There is no heat source in the restroom, and the toilet doors are closed to simulate the scenario when the restroom is in use. Geometric data for the model under investigation are listed in Table 2.  The major cause of poor air quality in restrooms is the malodor of feces and urine. The smell of feces and urine is primarily the smell of ammonia. In contrast, sewer malodor consists of mixed chemical substances, such as hydrogen sulfide, methyl mercaptan, trimethylamine, dimethyl disulfide, indole and methyl indole. This study focuses on the restroom flow field and the malodor concentration field. Because the irritating odor of ammonia is believed to be a major contributor to the offensive odor of human waste [5], the NH3 concentration in the malodor is analyzed. To simulate the very worst condition, the generation rate of unpleasant odors (represented by NH3) was assumed to be 0.3 L/min (0.2 g (NH3)/min) in this study. The pollution source area is set to 0.1 m × 0.1 m.
A modified odor removal efficiency (ORE) [2,16,17] was employed to express the ventilation performance of the whole restroom. A higher ORE indicates a lower concentration level, thus indicating better ventilation efficiency for odor removal. The ORE is defined as: is a southwesterly or southerly monsoon, while in winter it is northeasterly or northerly monsoon. Thus, the wind directions set in the CFD simulation include south and north winds. There is no heat source in the restroom, and the toilet doors are closed to simulate the scenario when the restroom is in use. Geometric data for the model under investigation are listed in Table 2.  The major cause of poor air quality in restrooms is the malodor of feces and urine. The smell of feces and urine is primarily the smell of ammonia. In contrast, sewer malodor consists of mixed chemical substances, such as hydrogen sulfide, methyl mercaptan, trimethylamine, dimethyl disulfide, indole and methyl indole. This study focuses on the restroom flow field and the malodor concentration field. Because the irritating odor of ammonia is believed to be a major contributor to the offensive odor of human waste [5], the NH3 concentration in the malodor is analyzed. To simulate the very worst condition, the generation rate of unpleasant odors (represented by NH3) was assumed to be 0.3 L/min (0.2 g (NH3)/min) in this study. The pollution source area is set to 0.1 m × 0.1 m.
A modified odor removal efficiency (ORE) [2,16,17] was employed to express the ventilation performance of the whole restroom. A higher ORE indicates a lower concentration level, thus indicating better ventilation efficiency for odor removal. The ORE is defined as: is a southwesterly or southerly monsoon, while in winter it is northeasterly or northerly monsoon. Thus, the wind directions set in the CFD simulation include south and north winds. There is no heat source in the restroom, and the toilet doors are closed to simulate the scenario when the restroom is in use. Geometric data for the model under investigation are listed in Table 2.  The major cause of poor air quality in restrooms is the malodor of feces and urine. The smell of feces and urine is primarily the smell of ammonia. In contrast, sewer malodor consists of mixed chemical substances, such as hydrogen sulfide, methyl mercaptan, trimethylamine, dimethyl disulfide, indole and methyl indole. This study focuses on the restroom flow field and the malodor concentration field. Because the irritating odor of ammonia is believed to be a major contributor to the offensive odor of human waste [5], the NH3 concentration in the malodor is analyzed. To simulate the very worst condition, the generation rate of unpleasant odors (represented by NH3) was assumed to be 0.3 L/min (0.2 g (NH3)/min) in this study. The pollution source area is set to 0.1 m × 0.1 m.
A modified odor removal efficiency (ORE) [2,16,17] was employed to express the ventilation performance of the whole restroom. A higher ORE indicates a lower concentration level, thus indicating better ventilation efficiency for odor removal. The ORE is defined as: is a southwesterly or southerly monsoon, while in winter it is northeasterly or northerly monsoon. Thus, the wind directions set in the CFD simulation include south and north winds. There is no heat source in the restroom, and the toilet doors are closed to simulate the scenario when the restroom is in use. Geometric data for the model under investigation are listed in Table 2.  The major cause of poor air quality in restrooms is the malodor of feces and urine. The smell of feces and urine is primarily the smell of ammonia. In contrast, sewer malodor consists of mixed chemical substances, such as hydrogen sulfide, methyl mercaptan, trimethylamine, dimethyl disulfide, indole and methyl indole. This study focuses on the restroom flow field and the malodor concentration field. Because the irritating odor of ammonia is believed to be a major contributor to the offensive odor of human waste [5], the NH3 concentration in the malodor is analyzed. To simulate the very worst condition, the generation rate of unpleasant odors (represented by NH3) was assumed to be 0.3 L/min (0.2 g (NH3)/min) in this study. The pollution source area is set to 0.1 m × 0.1 m.
A modified odor removal efficiency (ORE) [2,16,17] was employed to express the ventilation performance of the whole restroom. A higher ORE indicates a lower concentration level, thus indicating better ventilation efficiency for odor removal. The ORE is defined as: The major cause of poor air quality in restrooms is the malodor of feces and urine. The smell of feces and urine is primarily the smell of ammonia. In contrast, sewer malodor consists of mixed chemical substances, such as hydrogen sulfide, methyl mercaptan, trimethylamine, dimethyl disulfide, indole and methyl indole. This study focuses on the restroom flow field and the malodor concentration field. Because the irritating odor of ammonia is believed to be a major contributor to the offensive odor of human waste [5], the NH 3 concentration in the malodor is analyzed. To simulate the very worst condition, the generation rate of unpleasant odors (represented by NH 3 ) was assumed to be 0.3 L/min (0.2 g (NH 3 )/min) in this study. The pollution source area is set to 0.1 m × 0.1 m. A modified odor removal efficiency (ORE) [2,16,17] was employed to express the ventilation performance of the whole restroom. A higher ORE indicates a lower concentration level, thus indicating better ventilation efficiency for odor removal. The ORE is defined as: where C e is the odor concentration at the exhaust, C 0 is the indoor background concentration and C p is the average concentration at the height of the breathing zone.

Numerical Methods
Numerical simulations of the problem that is being investigated are performed via a finite volume method to solve the governing equations with the previously discussed boundary conditions ( Table 3). The calculation domain (50 m × 50 m × 3.2 m) is shown in Figure 3a. The commercial CFD code PHOENICS is used to simulate the airflow and NH 3 distributions. The governing equations solved by PHOENICS include a three-dimensional time-dependent incompressible Navier-Stokes equation, a time-independent convection diffusion equation and a k-ε turbulence equation. The formulations of these equations can be found in the PHOENICS manual [18] and in most CFD textbooks; thus, they are not provided here. The empirical turbulence coefficients for the k-ε turbulence equation are assigned as follows: σ k = 1.0, σ ε = 1.22, σ ε1 = 1.44, σ ε2 = 1.92 and Cµ = 0.09. These values are widely accepted in CFD k-ε models. To bridge the steep gradients of dependent variables near a solid surface, a general wall function is employed. Iterative calculation continues until a prescribed relative convergence of 10 −3 is satisfied for all field variables of this problem.
When testing the grid independence of a mesh domain, the NH 3 distribution at the user's squatting position (X = 0.925 m, Y = 1.2 m), which is based on different grid points, is used to calculate the deviation percentages and determine a suitable grid point system for the calculation (Figure 3b). Numerical simulation accuracy depends on the resolution of the computational mesh. A finer grid produces more accurate solutions. In this study, a grid system with approximately 131 × 113 × 52 (769,756) cells is used for numerical simulations. Each cell in the investigated restroom is about 0.1 m × 0.15 m × 0.06 m. An increase in the number of cells provides better information. However, such an increase is accompanied by a significant increase in computational resources.

Model Validation
In this study, a reduced-scale model of the investigated restroom shown in Figure 1 is constructed (Figure 4a). The material used for the model is 3-mm-thick gray hard cardboard and foam core board. The model scale is 1:45. The opening in the model is simply an opening without an installed window. A 4-inch fan is installed at the restroom entrance to simulate a south wind, and the airflow velocity in the model was measured by a multifunction measuring instrument (Testo 435-1) with an anemometer (Testo 0635 1535). There are seven measurement locations: at the two entrances (two locations), the window opening centers (two locations), the aisle centers (two locations) and the central toilet (one point). Each measurement location measures three heights: 1 cm, 3 cm and 5 cm. Next, CFD simulation with the same method mentioned in Section 2.2 is performed based on this reduced-scale model, and the simulation result is compared with the reduced-scale test result. As shown in Figure 4b, the difference between the CFD results and the experimental results is not significant. Thus, the reliability of the simulation results was confirmed.

Model Validation
In this study, a reduced-scale model of the investigated restroom shown in Figure 1 is constructed ( Figure 4a). The material used for the model is 3-mm-thick gray hard cardboard and foam core board. The model scale is 1:45. The opening in the model is simply an opening without an installed window. A 4-inch fan is installed at the restroom entrance to simulate a south wind, and the airflow velocity in the model was measured by a multifunction measuring instrument (Testo 435-1) with an anemometer (Testo 0635 1535). There are seven measurement locations: at the two entrances (two locations), the window opening centers (two locations), the aisle centers (two locations) and the central toilet (one point). Each measurement location measures three heights: 1 cm, 3 cm and 5 cm. Next, CFD simulation with the same method mentioned in Section 2.2 is performed based on this reduced-scale model, and the simulation result is compared with the reduced-scale test result. As shown in Figure 4b, the difference between the CFD results and the experimental results is not significant. Thus, the reliability of the simulation results was confirmed.

Model Validation
In this study, a reduced-scale model of the investigated restroom shown in Figure 1 is constructed ( Figure 4a). The material used for the model is 3-mm-thick gray hard cardboard and foam core board. The model scale is 1:45. The opening in the model is simply an opening without an installed window. A 4-inch fan is installed at the restroom entrance to simulate a south wind, and the airflow velocity in the model was measured by a multifunction measuring instrument (Testo 435-1) with an anemometer (Testo 0635 1535). There are seven measurement locations: at the two entrances (two locations), the window opening centers (two locations), the aisle centers (two locations) and the central toilet (one point). Each measurement location measures three heights: 1 cm, 3 cm and 5 cm. Next, CFD simulation with the same method mentioned in Section 2.2 is performed based on this reduced-scale model, and the simulation result is compared with the reduced-scale test result. As shown in Figure 4b, the difference between the CFD results and the experimental results is not significant. Thus, the reliability of the simulation results was confirmed.

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH 3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45 • and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right ( Figure 5a, symbol

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right ( Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom ( Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom ( Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH3 concentrations in the three individual urinal areas are low ( Figure  5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH3 concentration than in other locations.

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right ( Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom ( Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom ( Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH3 concentrations in the three individual urinal areas are low ( Figure  5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH3 concentration than in other locations. ). This flow becomes the main stream that drives indoor air circulation (

Case Study: the Effect of Wind
In Taiwan, the perennial w section, the indoor flow field directions are investigated using at the center of the exterior nort 45° and an outdoor wind speed Figure 5a shows the flow north wind flows from the left inclined window surface, outdo upper right (Figure 5a, symbol (right side of the diagram; ). T (). Figure 5b shows the flow f Under the main stream , three the toilet area. Their flow speeds the main stream, they flow to th circulation. This circulation flow toward the top left . Then, it m south (right side of diagram). Be pattern, the flow in space  a produces a higher NH3 concentr Figure 5d shows the flow f Driven by main stream A and af refer to the symbols on Figure through the urinal area, the NH 5e). Affected by the flow struc concentration than in other loca ). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c,

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c).

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c).

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c).

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom (Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH3 concentrations in the three individual urinal areas are low ( Figure  5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH3 concentration than in other locations.  an, the perennial wind direction pattern is south in summer and north in winter. In this indoor flow field and NH3 concentration distribution for different outdoor wind e investigated using a case study with the following settings: awning windows installed of the exterior north wall (Table 1c,  n Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this n, the indoor flow field and NH3 concentration distribution for different outdoor wind ions are investigated using a case study with the following settings: awning windows installed center of the exterior north wall (Table 1c,  ttern is south in summer and north in winter. In this entration distribution for different outdoor wind th the following settings: awning windows installed , window configuration W2), an opening angle θ of in the women's restroom (Figure 1c, section A). A gram toward the awning window. Guided by the ard the indoor ceiling at an angle pointing to the s along the ceiling to the exterior door in the south s the main stream that drives indoor air circulation area in the women's restroom (Figure 1c, section B). are formed. The clockwise circulations  form in to carry NH3 to the top. After these flows merge into n the south. Circulation  is a large-scale clockwise a from the main stream, carries away NH3 and flows ain stream and flows toward the exterior door in the n  is weak, and the circulation follows a particular  is weak and unable to carry away NH3, which ncentration in the other toilet areas (Figure 5c). l area in the men's restroom (Figure 1c

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom (Figure 1c is weak and unable to carry away NH 3 , which produces a higher NH 3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom (Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH 3 concentrations in the three individual urinal areas are low (Figure 5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH 3 concentration than in other locations.

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom (Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH3 concentrations in the three individual urinal areas are low ( Figure  5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH3 concentration than in other locations. The preceding analysis and Figure 6 reveal that the toilet area and the urinal area near the exterior wall have poor ventilation due to the flow field structure. Figure 6a shows that when users defecate in a squatting position, the NH3 concentration is high at the breathing zone height, in the area marked with a red dotted line (Z = 0.6 m). When men urinate in a standing position, Figure 6b shows that the NH3 concentration is high at the breathing zone height, in the area marked with a red dotted line (Z = 1.5 m).
The Annex I 109.03 of "Information notices on occupational diseases: a guide to diagnosis (European Commission, 2009)" [19] indicates that the odor threshold of NH3 is about 20 ppm; exposure levels of NH3 that surpass 50 ppm will result in immediate irritation to the nose and throat; exposure level to 250 ppm is bearable for 30-60 min; and exposure level to 300 ppm is considered to be immediately dangerous to life and health. It is good to define an acceptable level of NH3 from which the ventilation performance of each case could be evaluated. However, the recommended values above cannot be well applied in quasi-steady-state problems raised in this study that urinating or defecating is within a limited time-period. More observation and discussion are needed. Besides, the generation rate of unpleasant odors set in this study presents the very worst condition; if a referenced threshold level was used and linked to our simulations, the results would be misleading. Such constraints limit this study to a relative comparison among cases. The preceding analysis and Figure 6 reveal that the toilet area and the urinal area near the exterior wall have poor ventilation due to the flow field structure. Figure 6a shows that when users defecate in a squatting position, the NH 3 concentration is high at the breathing zone height, in the area marked with a red dotted line (Z = 0.6 m). When men urinate in a standing position, Figure 6b shows that the NH 3 concentration is high at the breathing zone height, in the area marked with a red dotted line (Z = 1.5 m).
The Annex I 109.03 of "Information notices on occupational diseases: a guide to diagnosis (European Commission, 2009)" [19] indicates that the odor threshold of NH 3 is about 20 ppm; exposure levels of NH 3 that surpass 50 ppm will result in immediate irritation to the nose and throat; exposure level to 250 ppm is bearable for 30-60 min; and exposure level to 300 ppm is considered to be immediately dangerous to life and health. It is good to define an acceptable level of NH 3 from which the ventilation performance of each case could be evaluated. However, the recommended values above cannot be well applied in quasi-steady-state problems raised in this study that urinating or defecating is within a limited time-period. More observation and discussion are needed. Besides, the generation rate of unpleasant odors set in this study presents the very worst condition; if a referenced threshold level was used and linked to our simulations, the results would be misleading. Such constraints limit this study to a relative comparison among cases.

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom (Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH3 concentrations in the three individual urinal areas are low ( Figure  5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH3 concentration than in other locations.

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. section, the indoor flow field and NH3 concentration distribution for different outdoor directions are investigated using a case study with the following settings: awning windows in at the center of the exterior north wall (Table 1c, window configuration W2), an opening ang 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section north wind flows from the left side of the diagram toward the awning window. Guided inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the (right side of the diagram; ). This flow becomes the main stream that drives indoor air circ (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, sec Under the main stream , three air circulations are formed. The clockwise circulations  f the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows mer the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clo circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and toward the top left . Then, it merges into the main stream and flows toward the exterior doo south (right side of diagram). Because circulation  is weak, and the circulation follows a pa pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom (Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH3 concentrations in the three individual urinal areas are low ( Figure  5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH3 concentration than in other locations.

(a)
) at the bottom zone.     The preceding analysis and Figure 8 show that as a result of the flow structure, the toilet areas and urinal area near the indoor partition wall have inferior ventilation. Figure 8a shows that when users defecate in a squatting position the NH3 concentration is high at the breathing zone height, in the area marked with a red dotted line (Z = 0.6 m). When men urinate in a standing position, the NH3 concentration is high at the breathing zone height, in the area marked with a red dotted line (Z = 1.5 m) (Figure 8b). A possible solution is to add a vertical guiding board to the ceiling in the area with inferior ventilation (red blocks in Figures 6 and 8). In this manner, part of the main stream flow is guided into an area with poor ventilation, and the NH3 is carried away. Although this issue is not the focus of this study, it warrants for future investigation.

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom (Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH3 concentrations in the three individual urinal areas are low ( Figure  5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH3 concentration than in other locations.

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom (Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH3 concentrations in the three individual urinal areas are low ( Figure  5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH3 concentration than in other locations.

(a)
is formed, which results in a low NH 3 concentration in the restroom (Figure 7c). Because the main stream is close to the ceiling and the flow speed is low, stagnant air and high NH 3 concentrations result in the spaces at the bottom 7 of 14 attern is south in summer and north in winter. In this centration distribution for different outdoor wind with the following settings: awning windows installed 1c, window configuration W2), an opening angle θ of le in the women's restroom (Figure 1c, section A). A iagram toward the awning window. Guided by the oward the indoor ceiling at an angle pointing to the ws along the ceiling to the exterior door in the south mes the main stream that drives indoor air circulation t area in the women's restroom (Figure 1c, section B). ns are formed. The clockwise circulations  form in h to carry NH3 to the top. After these flows merge into r in the south. Circulation  is a large-scale clockwise rea from the main stream, carries away NH3 and flows main stream and flows toward the exterior door in the ion  is weak, and the circulation follows a particular f  is weak and unable to carry away NH3, which concentration in the other toilet areas (Figure 5c). nal area in the men's restroom (Figure 1c Wind Direction ial wind direction pattern is south in summer and north in winter. In this field and NH3 concentration distribution for different outdoor wind using a case study with the following settings: awning windows installed r north wall (Table 1c, window configuration W2), an opening angle θ of peed of 1 m/s. flow field in the aisle in the women's restroom (Figure 1c, section A). A e left side of the diagram toward the awning window. Guided by the outdoor air flows toward the indoor ceiling at an angle pointing to the bol ). Then, it flows along the ceiling to the exterior door in the south ). This flow becomes the main stream that drives indoor air circulation low field in the toilet area in the women's restroom (Figure 1c, section B). three air circulations are formed. The clockwise circulations  form in peeds are fast enough to carry NH3 to the top. After these flows merge into to the exterior door in the south. Circulation  is a large-scale clockwise flows to the toilet area from the main stream, carries away NH3 and flows , it merges into the main stream and flows toward the exterior door in the m). Because circulation  is weak, and the circulation follows a particular  at the bottom of  is weak and unable to carry away NH3, which centration than the concentration in the other toilet areas (Figure 5c). flow field in the urinal area in the men's restroom (Figure 1c

Case Study: the Effect of Wind Direction
In Taiwan, the perennial wind direction pattern is south in summer and north in winter. In this section, the indoor flow field and NH3 concentration distribution for different outdoor wind directions are investigated using a case study with the following settings: awning windows installed at the center of the exterior north wall (Table 1c, window configuration W2), an opening angle θ of 45° and an outdoor wind speed of 1 m/s. Figure 5a shows the flow field in the aisle in the women's restroom (Figure 1c, section A). A north wind flows from the left side of the diagram toward the awning window. Guided by the inclined window surface, outdoor air flows toward the indoor ceiling at an angle pointing to the upper right (Figure 5a, symbol ). Then, it flows along the ceiling to the exterior door in the south (right side of the diagram; ). This flow becomes the main stream that drives indoor air circulation (). Figure 5b shows the flow field in the toilet area in the women's restroom (Figure 1c, section B). Under the main stream , three air circulations are formed. The clockwise circulations  form in the toilet area. Their flow speeds are fast enough to carry NH3 to the top. After these flows merge into the main stream, they flow to the exterior door in the south. Circulation  is a large-scale clockwise circulation. This circulation flows to the toilet area from the main stream, carries away NH3 and flows toward the top left . Then, it merges into the main stream and flows toward the exterior door in the south (right side of diagram). Because circulation  is weak, and the circulation follows a particular pattern, the flow in space  at the bottom of  is weak and unable to carry away NH3, which produces a higher NH3 concentration than the concentration in the other toilet areas (Figure 5c). Figure 5d shows the flow field in the urinal area in the men's restroom (Figure 1c, section C). Driven by main stream A and affected by the urinal partition board, the clockwise flows BCD (please refer to the symbols on Figure 5d) merge with the counter-clockwise flow E. As flow BCD flows through the urinal area, the NH3 concentrations in the three individual urinal areas are low ( Figure  5e). Affected by the flow structure, the air in space F stagnates, which results in a higher NH3 concentration than in other locations.

(a)
. Figure 7d shows the flow field in the urinal area in the men's restroom (Figure 1c, section C). Main stream A flows into the room via the south exterior door (right side of the diagram). Near the north wall (left side of the diagram), part of the flow flows out the window, but the majority of the flow flows downward along the wall surface (B). Affected by the partition board, flow B turns right and exhibits a pattern of horizontal flow (C). Then, it turns upward at the partition wall (D) and forms a major counter-clockwise circulation (BCD). This large circulation surrounds the urinal area, which results in a high NH 3 concentration in this area (Figure 7e).
The preceding analysis and Figure 8 show that as a result of the flow structure, the toilet areas and urinal area near the indoor partition wall have inferior ventilation. Figure 8a shows that when users defecate in a squatting position the NH 3 concentration is high at the breathing zone height, in the area marked with a red dotted line (Z = 0.6 m). When men urinate in a standing position, the NH 3 concentration is high at the breathing zone height, in the area marked with a red dotted line (Z = 1.5 m) (Figure 8b). A possible solution is to add a vertical guiding board to the ceiling in the area with inferior ventilation (red blocks in Figures 6 and 8). In this manner, part of the main stream flow is guided into an area with poor ventilation, and the NH 3 is carried away. Although this issue is not the focus of this study, it warrants for future investigation.

Findings and Design Recommendation
The previous section explains how the ventilation performance of each restroom is analyzed via flow pattern observation and the NH3 concentration distribution. Due to page length limitations, this paper cannot elaborate the restroom ventilation effectiveness of each window configuration. Instead, the modified odor removal efficiency (ORE) was used to investigate the ventilation effectiveness in each case. Figure 9 shows the ORE for the cases with 0.5 m/s south wind (commonly seen in spring/summer) and 2.0 m/s north wind (autumn/winter). This figure reveals that the window configuration (W1-W7) has a major impact on the ventilation of the restrooms; this impact significantly exceeds the effect of other parameters (wind speed, wind direction and opening angle). When window configuration W2 is employed, an arbitrary opening angle (30°-60°) under the two wind conditions is acceptable; opening angle 45° is proposed. When window configuration W5 is adopted, the 30° and 60° opening angle configurations are recommended for spring/summer and autumn/winter respectively (marked in light yellow and light blue in Figure 9).

Findings and Design Recommendation
The previous section explains how the ventilation performance of each restroom is analyzed via flow pattern observation and the NH 3 concentration distribution. Due to page length limitations, this paper cannot elaborate the restroom ventilation effectiveness of each window configuration. Instead, the modified odor removal efficiency (ORE) was used to investigate the ventilation effectiveness in each case. Figure 9 shows the ORE for the cases with 0.5 m/s south wind (commonly seen in spring/summer) and 2.0 m/s north wind (autumn/winter). This figure reveals that the window configuration (W1-W7) has a major impact on the ventilation of the restrooms; this impact significantly exceeds the effect of other parameters (wind speed, wind direction and opening angle). When window configuration W2 is employed, an arbitrary opening angle (30 • -60 • ) under the two wind conditions is acceptable; opening angle 45 • is proposed. When window configuration W5 is adopted, the 30 • and 60 • opening angle configurations are recommended for spring/summer and autumn/winter respectively (marked in light yellow and light blue in Figure 9).

Findings and Design Recommendation
The previous section explains how the ventilation performance of each restroom is analyzed via flow pattern observation and the NH3 concentration distribution. Due to page length limitations, this paper cannot elaborate the restroom ventilation effectiveness of each window configuration. Instead, the modified odor removal efficiency (ORE) was used to investigate the ventilation effectiveness in each case. Figure 9 shows the ORE for the cases with 0.5 m/s south wind (commonly seen in spring/summer) and 2.0 m/s north wind (autumn/winter). This figure reveals that the window configuration (W1-W7) has a major impact on the ventilation of the restrooms; this impact significantly exceeds the effect of other parameters (wind speed, wind direction and opening angle). When window configuration W2 is employed, an arbitrary opening angle (30°-60°) under the two wind conditions is acceptable; opening angle 45° is proposed. When window configuration W5 is adopted, the 30° and 60° opening angle configurations are recommended for spring/summer and autumn/winter respectively (marked in light yellow and light blue in Figure 9).  Design recommendations for the installation quantity and position of awning windows are listed in Table 4. K-12 school restrooms could adopt window configuration W2 for the north walls with an opening angle of 45 • for all seasons. Window configuration W5 is also recommended. However, here, different seasons require different opening angles. In spring and summer with low-speed wind, the window should be opened 30 • . Such a configuration could also prevent rain from entering the room during the rainy season in summer. In autumn and winter with stronger wind, the window should be opened 60 • . However, the cost of the W5 window configuration is high. In addition, some of the windows are located in high positions and are thus difficult to open. Therefore, window configuration W5 is not our first recommendation. Design recommendations for the installation quantity and position of awning windows are listed in Table 4. K-12 school restrooms could adopt window configuration W2 for the north walls with an opening angle of 45° for all seasons. Window configuration W5 is also recommended. However, here, different seasons require different opening angles. In spring and summer with lowspeed wind, the window should be opened 30°. Such a configuration could also prevent rain from entering the room during the rainy season in summer. In autumn and winter with stronger wind, the window should be opened 60°. However, the cost of the W5 window configuration is high. In addition, some of the windows are located in high positions and are thus difficult to open. Therefore, window configuration W5 is not our first recommendation.

Conclusions
In this study, the restrooms in K-12 public schools in Taiwan are selected as the study subjects. The effect of awning window quantity and installation position on airflow pattern and air contaminant (NH3) distribution in the representative restroom is analyzed using CFD. The research findings are summarized as follows.
1. In autumn and winter, the north monsoon wind flows into the awning window from the north.
Guided by the inclined window surface, it flows toward the indoor ceiling at an angle that is oriented upwards. It then flows to the south exterior door along the ceiling. In spring and summer, a south monsoon occurs, and a reverse indoor flow pattern is observed. This northsouth flow becomes the main stream that drives restroom indoor circulation at the bottom (and along both sides). 2. In the toilet areas in the men's and women's restrooms and the urinal area in the men's restroom, if the air flow into these areas forms a circulation pattern in the main stream and the flow speed is sufficiently high, the air circulation will carry air pollutants from the areas to the top of the space, where they flow out of the restroom after merging with the main stream. If these conditions are not satisfied, the air in the areas will stagnate, and the air pollutant concentration will be high. Under a south wind in summer, the flow at the bottom of the main stream is affected by the urinal partition board and partition wall and forms a large circulation surrounding the urinal area. This circulation causes poor ventilation in this area. 3. Based on the ventilation performance analysis using the modified odor removal efficiency (ORE), we suggest that K-12 school restrooms use window configuration W2 in their north walls (as shown in Table 4). The opening angle should be set to 45° for all seasons.
The modeling and simulation results are limited to the urban and building morphologies chosen with the specific wind environments. At other conditions, recommendations given in the paper may Design recommendations for the installation quantity and position of awning windows are listed in Table 4. K-12 school restrooms could adopt window configuration W2 for the north walls with an opening angle of 45° for all seasons. Window configuration W5 is also recommended. However, here, different seasons require different opening angles. In spring and summer with lowspeed wind, the window should be opened 30°. Such a configuration could also prevent rain from entering the room during the rainy season in summer. In autumn and winter with stronger wind, the window should be opened 60°. However, the cost of the W5 window configuration is high. In addition, some of the windows are located in high positions and are thus difficult to open. Therefore, window configuration W5 is not our first recommendation.

Conclusions
In this study, the restrooms in K-12 public schools in Taiwan are selected as the study subjects. The effect of awning window quantity and installation position on airflow pattern and air contaminant (NH3) distribution in the representative restroom is analyzed using CFD. The research findings are summarized as follows.
1. In autumn and winter, the north monsoon wind flows into the awning window from the north.
Guided by the inclined window surface, it flows toward the indoor ceiling at an angle that is oriented upwards. It then flows to the south exterior door along the ceiling. In spring and summer, a south monsoon occurs, and a reverse indoor flow pattern is observed. This northsouth flow becomes the main stream that drives restroom indoor circulation at the bottom (and along both sides). 2. In the toilet areas in the men's and women's restrooms and the urinal area in the men's restroom, if the air flow into these areas forms a circulation pattern in the main stream and the flow speed is sufficiently high, the air circulation will carry air pollutants from the areas to the top of the space, where they flow out of the restroom after merging with the main stream. If these conditions are not satisfied, the air in the areas will stagnate, and the air pollutant concentration will be high. Under a south wind in summer, the flow at the bottom of the main stream is affected by the urinal partition board and partition wall and forms a large circulation surrounding the urinal area. This circulation causes poor ventilation in this area. 3. Based on the ventilation performance analysis using the modified odor removal efficiency (ORE), we suggest that K-12 school restrooms use window configuration W2 in their north walls (as shown in Table 4). The opening angle should be set to 45° for all seasons.
The modeling and simulation results are limited to the urban and building morphologies chosen with the specific wind environments. At other conditions, recommendations given in the paper may Second recommendation W5 Design recommendations for the installation quantity and position of awning windows are listed in Table 4. K-12 school restrooms could adopt window configuration W2 for the north walls with an opening angle of 45° for all seasons. Window configuration W5 is also recommended. However, here, different seasons require different opening angles. In spring and summer with lowspeed wind, the window should be opened 30°. Such a configuration could also prevent rain from entering the room during the rainy season in summer. In autumn and winter with stronger wind, the window should be opened 60°. However, the cost of the W5 window configuration is high. In addition, some of the windows are located in high positions and are thus difficult to open. Therefore, window configuration W5 is not our first recommendation. Autumn and winter 60°

Conclusions
In this study, the restrooms in K-12 public schools in Taiwan are selected as the study subjects. The effect of awning window quantity and installation position on airflow pattern and air contaminant (NH3) distribution in the representative restroom is analyzed using CFD. The research findings are summarized as follows.
1. In autumn and winter, the north monsoon wind flows into the awning window from the north.
Guided by the inclined window surface, it flows toward the indoor ceiling at an angle that is oriented upwards. It then flows to the south exterior door along the ceiling. In spring and summer, a south monsoon occurs, and a reverse indoor flow pattern is observed. This northsouth flow becomes the main stream that drives restroom indoor circulation at the bottom (and along both sides). 2. In the toilet areas in the men's and women's restrooms and the urinal area in the men's restroom, if the air flow into these areas forms a circulation pattern in the main stream and the flow speed is sufficiently high, the air circulation will carry air pollutants from the areas to the top of the space, where they flow out of the restroom after merging with the main stream. If these conditions are not satisfied, the air in the areas will stagnate, and the air pollutant concentration will be high. Under a south wind in summer, the flow at the bottom of the main stream is affected by the urinal partition board and partition wall and forms a large circulation surrounding the urinal area. This circulation causes poor ventilation in this area. 3. Based on the ventilation performance analysis using the modified odor removal efficiency (ORE), we suggest that K-12 school restrooms use window configuration W2 in their north walls (as shown in Table 4). The opening angle should be set to 45° for all seasons.
The modeling and simulation results are limited to the urban and building morphologies chosen with the specific wind environments. At other conditions, recommendations given in the paper may (Window number: 8) Design recommendations for the installation quantity and position of awning windows are listed in Table 4. K-12 school restrooms could adopt window configuration W2 for the north walls with an opening angle of 45° for all seasons. Window configuration W5 is also recommended. However, here, different seasons require different opening angles. In spring and summer with lowspeed wind, the window should be opened 30°. Such a configuration could also prevent rain from entering the room during the rainy season in summer. In autumn and winter with stronger wind, the window should be opened 60°. However, the cost of the W5 window configuration is high. In addition, some of the windows are located in high positions and are thus difficult to open. Therefore, window configuration W5 is not our first recommendation. Autumn and winter 60°

Conclusions
In this study, the restrooms in K-12 public schools in Taiwan are selected as the study subjects. The effect of awning window quantity and installation position on airflow pattern and air contaminant (NH3) distribution in the representative restroom is analyzed using CFD. The research findings are summarized as follows.
1. In autumn and winter, the north monsoon wind flows into the awning window from the north.
Guided by the inclined window surface, it flows toward the indoor ceiling at an angle that is oriented upwards. It then flows to the south exterior door along the ceiling. In spring and summer, a south monsoon occurs, and a reverse indoor flow pattern is observed. This northsouth flow becomes the main stream that drives restroom indoor circulation at the bottom (and along both sides). 2. In the toilet areas in the men's and women's restrooms and the urinal area in the men's restroom, if the air flow into these areas forms a circulation pattern in the main stream and the flow speed is sufficiently high, the air circulation will carry air pollutants from the areas to the top of the space, where they flow out of the restroom after merging with the main stream. If these conditions are not satisfied, the air in the areas will stagnate, and the air pollutant concentration will be high. Under a south wind in summer, the flow at the bottom of the main stream is affected by the urinal partition board and partition wall and forms a large circulation surrounding the urinal area. This circulation causes poor ventilation in this area. 3. Based on the ventilation performance analysis using the modified odor removal efficiency (ORE), we suggest that K-12 school restrooms use window configuration W2 in their north walls (as shown in Table 4). The opening angle should be set to 45° for all seasons.
The modeling and simulation results are limited to the urban and building morphologies chosen with the specific wind environments. At other conditions, recommendations given in the paper may Autumn and winter Design recommendations for the installation quantity and position of awning windows are listed in Table 4. K-12 school restrooms could adopt window configuration W2 for the north walls with an opening angle of 45° for all seasons. Window configuration W5 is also recommended. However, here, different seasons require different opening angles. In spring and summer with lowspeed wind, the window should be opened 30°. Such a configuration could also prevent rain from entering the room during the rainy season in summer. In autumn and winter with stronger wind, the window should be opened 60°. However, the cost of the W5 window configuration is high. In addition, some of the windows are located in high positions and are thus difficult to open. Therefore, window configuration W5 is not our first recommendation. Autumn and winter 60°

Conclusions
In this study, the restrooms in K-12 public schools in Taiwan are selected as the study subjects. The effect of awning window quantity and installation position on airflow pattern and air contaminant (NH3) distribution in the representative restroom is analyzed using CFD. The research findings are summarized as follows.
1. In autumn and winter, the north monsoon wind flows into the awning window from the north.
Guided by the inclined window surface, it flows toward the indoor ceiling at an angle that is oriented upwards. It then flows to the south exterior door along the ceiling. In spring and summer, a south monsoon occurs, and a reverse indoor flow pattern is observed. This northsouth flow becomes the main stream that drives restroom indoor circulation at the bottom (and along both sides). 2. In the toilet areas in the men's and women's restrooms and the urinal area in the men's restroom, if the air flow into these areas forms a circulation pattern in the main stream and the flow speed is sufficiently high, the air circulation will carry air pollutants from the areas to the top of the space, where they flow out of the restroom after merging with the main stream. If these conditions are not satisfied, the air in the areas will stagnate, and the air pollutant concentration will be high. Under a south wind in summer, the flow at the bottom of the main stream is affected by the urinal partition board and partition wall and forms a large circulation surrounding the urinal area. This circulation causes poor ventilation in this area. 3. Based on the ventilation performance analysis using the modified odor removal efficiency (ORE), we suggest that K-12 school restrooms use window configuration W2 in their north walls (as shown in Table 4). The opening angle should be set to 45° for all seasons.
The modeling and simulation results are limited to the urban and building morphologies chosen with the specific wind environments. At other conditions, recommendations given in the paper may

Conclusions
In this study, the restrooms in K-12 public schools in Taiwan are selected as the study subjects. The effect of awning window quantity and installation position on airflow pattern and air contaminant (NH 3 ) distribution in the representative restroom is analyzed using CFD. The research findings are summarized as follows.

1.
In autumn and winter, the north monsoon wind flows into the awning window from the north. Guided by the inclined window surface, it flows toward the indoor ceiling at an angle that is oriented upwards. It then flows to the south exterior door along the ceiling. In spring and summer, a south monsoon occurs, and a reverse indoor flow pattern is observed. This north-south flow becomes the main stream that drives restroom indoor circulation at the bottom (and along both sides).

2.
In the toilet areas in the men's and women's restrooms and the urinal area in the men's restroom, if the air flow into these areas forms a circulation pattern in the main stream and the flow speed is sufficiently high, the air circulation will carry air pollutants from the areas to the top of the space, where they flow out of the restroom after merging with the main stream. If these conditions are not satisfied, the air in the areas will stagnate, and the air pollutant concentration will be high. Under a south wind in summer, the flow at the bottom of the main stream is affected by the urinal partition board and partition wall and forms a large circulation surrounding the urinal area. This circulation causes poor ventilation in this area.

3.
Based on the ventilation performance analysis using the modified odor removal efficiency (ORE), we suggest that K-12 school restrooms use window configuration W2 in their north walls (as shown in Table 4). The opening angle should be set to 45 • for all seasons. The modeling and simulation results are limited to the urban and building morphologies chosen with the specific wind environments. At other conditions, recommendations given in the paper may not be applicable. The surroundings would greatly affect the magnitude and direction of the approaching wind. The interior partition design and layout would also affect the ventilation performance. Although investigation on other themes (surroundings, outside environments, interior partitioning, buoyancy effect, link to IAQ studies, etc.) are not what we are exploring in this study but is worthy of further consideration.

Conflicts of Interest:
The authors declare that there is no conflict of interest.