Wind Catcher Cooling Performance Including Heat Loads: An Experimental Study

Abstract

This study experimentally investigates the cooling performance of a single-opening wind catcher model under varying orientations and wind speeds. The wind catcher was connected to a horizontal cavity representing an indoor space, with a rear outlet simulating a window opening. Electric resistors were installed at the catcher shaft and in the middle of the cavity length to simulate the building’s heat loads. Experiments were conducted in a wind tunnel, where K-type thermocouples were employed to record temperature variations for both closed and open cavity ends. Five wind speeds (4–9 m/s) and five orientations (0–180°) were examined. Under the closed-cavity configuration, the maximum temperature reduction (cooling) of 4 °C occurred at an orientation of 180°, at which the catcher opening was positioned on the leeward side. This orientation created a low-pressure region at the catcher’s inlet, located within the wake of the model, which, combined with a favorable vertical temperature gradient, enhanced suction-driven cooling. In the open-cavity configuration, cooling was observed for all orientations and wind speeds. The greatest temperature reduction of 6 °C occurred at the 180° orientation, whereas other orientations produced lower temperatures changes, down to 2 °C.

1. Introduction

Efficient heating and cooling of buildings is of paramount importance for contemporary societies, as it directly impacts energy consumption, environmental sustainability, and economic costs. With buildings accounting for a significant portion of global energy use, optimizing thermal performance has become a central focus in both engineering and architectural design. In an extensive review by Chenari et al. [1] various energy-efficient methods are reported. In an effort to reduce energy consumption, traditional methods are being revisited, such as the so-called Arabic wind catchers, which have been used for centuries in Middle East to passively cool building interiors. Arabic wind catchers consist of a vertical shaft with openings at the top, designed to capture prevailing winds. These openings allow air to flow into the building, providing both ventilation and cooling. The air circulates in the interior, lowering temperatures without the need for mechanical cooling systems. Passive cooling in wind catchers is achieved through a combination of a vertical temperature gradient, created mainly by solar heat, and the pressure spatial differences generated by the blowing winds. The temperature variation causes air movement, with hot air masses rising, and cooler air being directed downward. Meanwhile, wind pressure helps draw cooler air into the building or expelling interior air when suction pressures prevail at the openings, promoting natural ventilation. This combination of forces significantly reduces the need for mechanical cooling, providing an energy-efficient solution. Egyptian architect Hassan Fathy revitalized the traditional wind catcher, integrating it into modern architectural practice as an effective passive cooling strategy for hot arid climates [2]. Comprehensive reviews of wind catcher applications are found in Saadatian et al. [3] and Hughes et al. [4], whereas the review by Jomehzadeh et al. [5] focuses on the performance of wind catchers with regard to indoor air quality.

The cooling performance of a wind catcher is influenced by multiple parameters, including the shaft height, the size, number, and orientation of its openings, and its alignment with respect to the prevailing wind direction. As highlighted in the review work by Khan et al. [6], the ventilation effectiveness of a building is governed not solely by the wind catcher, but by the combined influence of the wind catcher and the building’s window openings. Additionally, the temperature distribution of the building walls due to solar heat, the thermal properties of the building materials, and the temperature and pressure differences between the interior and exterior of the building all play crucial roles in determining the effectiveness of the wind catcher. In the numerical work of Cook [7], it is shown how the buoyancy forces due to vertical temperature gradients are opposed by the wind forces, influencing the indoor air motion accordingly.

Many publications appear in the available literature, either experimental or computational, exploring various parameters related to the performance of wind catchers. For example, based on wind tunnel experiments, it was found (Montazeri & Azizian [8]) that the internal airflow rate of a one-sided wind catcher attached to a building with an opening at its back side, is maximized for a zero-degree angle relative to the prevailing wind direction. Increasing the wind angle, the flow rate is reduced up to an angle of 68° at which the flow is minimized, while it changes direction for higher angles. In the latter case, the opening of the catcher becomes a flow outlet due to the prevailing low pressures in the wake of the wind catcher. The same authors in another publication (Montazeri & Azizian [9]), examining a model of a two-sided wind catcher model, found the internal flow rate to maximize at an angle of 90°, being 20% higher than for the zero-angle orientation angle. In Montazeri [10], the internal flow rate of multi-opening wind catchers was investigated both numerically and experimentally. The study found that increasing the number of openings reduces the sensitivity of the flow rate to wind direction. Based on the climatic conditions in Jordan, Badran [11] proposed a shorter wind catcher compared to traditional designs, suggesting that a 4 m tall tower with a square cross-section of 0.57 m per side can lower the indoor temperature from 36 °C to 25 °C. Dehghani-Sanij et al. [12] proposed a wind catcher with a rotating top, which can be aligned with the prevailing wind direction as well as a solar chimney at low wind speed sites. In Alsailani et al. [13], the focus is on the maximization of the internal flow of a wind catcher, proposing several shapes and guide vanes of the catcher’s inlets for the reduction in the flow recirculation zone at the catcher bend duct. In Chohan et al. [14], a traditional UAE (United Arab Emirates) two-floor house (Al Zarouni House) was computationally examined, including two X-blade wind catchers, concluding that they perform best from October to March, causing a maximum reduction in the interior temperature of 7 °C. Recently, the influence of the wind catchers in the natural ventilation of a two-story house was computationally examined, including various openings (Tantasavasdi et al. [15]). In a computational work by Foroozesh et al. [16], a one-sided wind catcher is examined, in which water droplets are sprayed in order to enhance the cooling performance, achieving a maximum interior temperature reduction of 17.4 °C. The same work includes an extensive list of relevant studies, providing a valuable source of information on wind catchers including water evaporation as an effective means for internal temperature reduction.

In Bahadori’s work [17], a system comprising several wind towers, a domed roof, and a basement water tank was examined. In this configuration, incoming air is cooled as it passes over the free surface of the water tank and then exits near the top of the domed roof, where a local pressure minimum exists, generated by the accelerated flow of the external air over the curved surface. In a subsequent study (Bahadori [18]), the same author proposed an alternative wind catcher design in which evaporative cooling is achieved by spraying water into the catcher’s shaft through a series of clay conduits. Later, in an experimental investigation (Bahadori et al. [19]), the cooling performance of traditional wind catchers was found to be inferior to two modified configurations: one employing wetted cloth curtains suspended along the entire height of the tower, and another incorporating wetted pads placed directly at the air inlets.

Recent research has increasingly focused on hybrid passive ventilation systems, which integrate wind-driven airflow with components such as solar chimneys, wind catchers, and evaporative cooling units. Purely wind-driven systems fail on calm days, and purely buoyancy-driven systems are weak at night or on overcast days when solar heating is significantly reduced, as the temperature difference between indoor and outdoor air is minimized. A hybrid system ensures continuous and stable ventilation by having one force (buoyancy force) available to compensate when the other is weak (pressure force). For instance, when wind is low, solar chimney maintains the airflow rate as it is shown in the numerical work of Yue et al. [20].

In a recent comprehensive review, Li et al. [21] discussed the integration of wind catchers with earth-air heat exchangers, heat pipes, and phase-change materials (PCMs). For instance, in Egypt, Mourad et al. [22] combined a wind catcher with a geothermal heat exchanger; in this configuration, air from the wind catcher shaft is routed through an underground channel to be cooled before entering the building. Similarly, Sakhri et al. [23] investigated a system where air is supplied via a wind catcher while outdoor air simultaneously enters the interior through a 60 m long horizontal tube buried 1.5 m underground. This setup increased thermal comfort levels by up to 50%, particularly when windows remained open. Furthermore, the influence of PCM on the performance of a two-sided wind catcher was experimentally examined by Abdo et al. [24]. In their study, the PCM was integrated into both the walls of the acrylic test chamber and the inlet duct of the wind catcher. Their findings demonstrated a reduction in the chamber temperature of approximately 3 °C. While further research is required to optimize these systems, the integration of PCM demonstrates significant potential for enhancing the thermal performance of wind catchers. In another approach, horizontal heat pipes installed in the vertical shaft of a wind catcher enhance heat exchange between the incoming and outgoing air streams. In summer, they help cool the incoming air before it enters the building, whereas in winter, they transfer heat from the exhaust air to the fresh air, warming it prior to entry. This idea was examined by Mahon et al. [25] both experimentally and numerically, proving that this passive system can reduce the building’s heating and cooling demand.

Investigations of wind catchers have traditionally focused on airflow behavior under isothermal conditions, primarily quantifying velocity fields, pressure distributions, and ventilation rates, while typically neglecting internal heat loads. However, omitting heat loads alters the governing physics of the problem since in the absence of temperature gradients, buoyancy forces are absent and indoor air motion becomes purely wind-driven. However, in reality, internal heat gains give rise to spatial temperature gradients that induce buoyancy-driven flow, modify internal pressure distributions, and interact with wind-induced ventilation, resulting in mixed-mode behavior that cannot be captured under isothermal assumptions. The present study addresses this limitation by explicitly incorporating internal heat loads into the evaluation of wind catcher performance. More particularly, the model was placed within the test section of an open-circuit wind tunnel, and its cooling performance was evaluated through temperature measurements taken by thermocouples. The study examined five different wind speeds, five orientations, and two configurations for the back-end of the wind catcher cavity, namely open and closed, simulating fifty cases in total. It was found that the catcher’s cooling performance depends on the vertical temperature gradient within its shaft as well as on the pressure distribution at the openings, which varies with wind direction. Notably, the cooling performance increased when the wind catcher’s opening was positioned on the leeward side of the catcher’s shaft, due to both a suction-driven flow mechanism and a favorable vertical temperature gradient. To the best of our knowledge, this behavior has not been reported previously, most likely because existing experimental studies do not incorporate internal heat loads and therefore do not capture the coupled wind–buoyancy mechanism observed here.

2. Materials and Methods

A model of a wind catcher was constructed from wood due to its favorable thermal insulation properties, ensuring the building envelope functioned as an effective thermal insulator. This minimized heat conduction through the walls, allowing the thermal influence of the external wind tunnel airflow to be neglected. Consequently, the temperature variations recorded by the thermocouples can be attributed solely to the air exchange driven by the wind catcher. Furthermore, the low thermal expansion coefficient of wood ensured that the internal geometry of the wind catcher remained rigid and dimensionally stable throughout the testing procedure. Finally, the model was treated with a moisture-resistant coating to prevent hygroscopic expansion and maintain surface integrity. The model consisted of a vertical shaft 120 mm in height with an internal cross-section of 38 mm $\times$ 38 mm. One of the four vertical faces of the shaft had an opening, measuring 30 mm (horizontal) $\times$ 35 mm (vertical). The shaft was connected to a horizontal cavity, also made of wood, representing an interior space. The cavity was 200 mm in length, with a cross-section of 88 mm (horizontal) $\times$ 68 mm (vertical), including an opening at its end, simulating a window, measuring 70 mm (horizontal) $\times$ 60 mm (vertical). The walls of the model were 6 mm thick. Figure 1 illustrates the geometric details of the model, including a three-dimensional schematic showing the wind interception at the front and back of the wind catcher. The chosen dimensions of the model are related to a real building through a scale factor. Namely, if the typical height of a residential building is considered to be 4 m ([26], p. 17), this size is 58.8 times larger compared to the 68 mm height of the model cavity. Therefore, assuming a scale factor of 58.8, the corresponding height of a real wind catcher would be 58.8 $\times$ 0.12 = 7.05 m, the side of its square cross-section would be 58.8 $\times$ 0.038 = 2.23 m, and its opening 1.76 m (horizontal) by 2.05 m (vertical). By comparison, in Bahadori et al. [19], a full scale wind catcher is examined with a net catcher height of 8 m, a cross-section of 1 m by 1 m, and an opening of 1.5 m, whereas in Ghadiri & Ibrahim [27], a 5 m tall catcher is studied with a 1.5 m side square cross-section and top openings of 1.5 m by 1.5 m.

Six flexible polyamide resistors (50 mm by 25 mm, R = 22 Ω each) were employed to simulate the building’s internal heat load, distributed between the vertical shaft and the horizontal cavity. In the vertical shaft (Figure 2a), three resistors were mounted 5 mm below the inlet, oriented with their longitudinal axes vertical. The photo was taken from the base of the shaft. One resistor was affixed to the wall containing the inlet opening, while the other two were placed on the adjacent side walls. The fourth wall was left unheated, assuming it was shielded from direct thermal exposure by the building structure. The remaining three resistors were installed in the horizontal cavity on the two vertical side walls and the ceiling. This configuration assumes these surfaces are exposed to solar radiation, while the floor is considered adiabatic (thermally insulated). These resistors were oriented with their longer dimension parallel to the cavity’s longitudinal axis, centered on their respective walls (Figure 2b). It is noted that the resistor at the center of Figure 2b is attached to the ceiling of the cavity model (a close up image is shown in Figure 2c).

The resistors were connected to an adjustable AC voltage source (Variac), and the voltage was maintained at a constant V = 10.5 V. Consequently, the total heat load (Q) applied to the interior of the structure was calculated as follows:

$$\mathrm{Q}=6{\mathrm{V}}^2/\mathrm{R}=6\times {10.5}^2/22=30.06 \mathrm{W}$$

(1)

The applied voltage was chosen to ensure that the resistors’ temperature did not exceed 90 °C, thereby preventing any damage to the model, while allowing spatial temperature variations to be detected by the temperature sensors. Based on the model’s total internal surface area of 792 cm², the heat load per unit internal wall area is calculated as 30.06/0.0792 ≈ 380 W/m². This value is higher than the typical heat loads reported for residential buildings in Bahrain. For example, Salem et al. [28] analyzed a two-story building with a floor area of 127 m² and windows covering 3.2% of the wall surface, in six Middle Eastern cities, including the capital of Bahrain. Their results indicate a mean heat load in the hottest month of July of approximately 100 W/m² for an indoor temperature of 22 °C (based on Figures 4 and 5 in [28]). It has to be stressed that in wind-tunnel model testing, exact thermal similitude cannot be achieved because dimensionless parameters such as the Grashof (Gr) and Richardson (Ri) numbers, which govern the Nusselt (Nu) number, scale strongly with the characteristic length. However, in natural convection, Nu depends on Gr by a power-law correlation, where the exponent typically varies between 1/4 and 1/3 depending on the flow regime (see [29]) and is also proportional to the characteristic length. Consequently, the convective heat transfer coefficient is weakly dependent on both geometric scaling and the wall-to-air temperature difference if natural convection is considered. Moreover, as discussed in paragraph 4, the calculated heat transfer coefficient is comparable to values reported in the literature.

The temperature of the resistor surface was measured with a digital infrared thermometer (MESTEK, IR02C), by manually scanning its surface, varying between 80 °C and 100 °C. The temperature in the model was measured using two K-type thermocouples, one positioned at the entrance of the vertical shaft (denoted as T₁) on the vertical wall opposite to the catcher’s opening central point, and the other, denoted as T₂, in the middle of the cavity at the bottom (floor) side, 20 mm downstream from the resistors. Both thermocouples were connected to a portable digital thermometer (Perfect Prime TC9815), which provided temperature readings with a resolution of 0.1 °C. To verify the reliability of the recorded data, a sensitivity test was performed by moving the temperature measurement point from the bottom wall to the mid-height of the cavity. At the examined wind speed of 9.04 m/s, the results remained qualitatively consistent, confirming that the measurement location did not alter the comparative conclusions of the study.

To account for the spatial temperature distribution within the wind catcher, five thermocouples were also installed across a vertical cross-section, positioned 20 mm downstream from the resistors toward the cavity exit. The locations of these sensors are shown in Figure 3a, in which the dimensions of 70 mm (width) by 60 mm (height) refer to the cavity window size. The thermocouples are shown in Figure 3b, and they are indexed with the numbers 1 through 5. Positions 1 and 2 comprise the upper horizontal row (close to the cavity ceiling) with number 1 being on the left side, position 3 marks the geometric center, and positions 4 and 5 form the lower horizontal row (close to the floor).

The wind catcher model was placed in the test section of a subsonic, open-circuit wind tunnel at Bahrain Polytechnic, with a cross-sectional area of 305 mm $\times$ 305 mm with a minimum reliable operating velocity of approximately 3 m/s. To ensure stable flow conditions, measurements were initiated at a wind speed of 4 m/s. Although this velocity is higher than typical low outdoor wind speeds observed near the ground in many urban environments, it ensures stable operation, a uniform velocity distribution, and acceptable turbulence intensity within the test section. Moreover, as the wind catcher shaft opening is typically located above the building roof level, the corresponding wind speeds are generally higher than 4 m/s, as reported in field measurements from urban environments (e.g., [30]). The velocity boundary layer over the floor of the wind tunnel, measured at the mid-length of the test section where the model was placed, is shown in Figure 4. Measurements were taken using a Pitot–static tube with an external diameter of 2 mm at the highest examined free-stream speed of 9.04 m/s. The boundary layer thickness for the lowest wind speed (4 m/s) is estimated at 46.5 mm, calculated using the 1/7th power law for turbulent flow. This is comparable to the 40 mm thickness observed at 9.04 m/s. Consequently, the air velocity reaches its free-stream value within 46 mm from the tunnel’s bottom surface in all examined cases. Given that the window opening extends vertically from 10 mm to 70 mm, and the center of the catcher’s opening is positioned at 176.5 mm (see Figure 2), the vertical variation in inlet velocity across these openings is mild.

Five different orientations of the model relative to the incoming free stream were investigated—0° (head-on wind, normal to the catcher’s opening, see also Figure 1), 45°, 90°, 135°, and 180° (opening in the lee side of the catcher’s shaft)—and five free-stream velocities: 4.04 m/s, 5.71 m/s, 8.08 m/s, and 9.04 m/s. The airflow was regulated by adjusting the rotational speed of the wind tunnel’s fan motor and was measured using a Pitot-Static tube located at the entrance of the test section. The dynamic pressure fluctuations of the free stream were within ±1% of the mean and the free stream temperature during testing was 19 °C, with a maximum variation of 0.5 °C.

The blockage ratio (projected model area to the cross sectional area of the wind tunnel) took a maximum value of 23.66% (at 90° model orientation), causing an increase in the free stream velocity of 0.25 $\times$ 0.2366 = 5.9% according to Barlow et al. [31] (p. 374, Equation 10.22).

For each combination of model orientation and free-stream velocity, two configurations were tested—one with the cavity at its back side open and the other being closed—simulating the case of a window at the back side of a building being open and closed, respectively. In total, 50 tests were conducted. For each case, the resistors were connected to the power source, and once the two temperatures, T₁ (at the catcher’s opening) and T₂ (at the middle of the cavity), were at 42 °C and 26 °C, respectively, with a deviation of 0.5 °C, the wind tunnel was turned on. After two minutes of airflow, the two temperatures were recorded. This approach enabled the assessment of the wind catcher’s cooling performance by evaluating the temperature change at the floor level (final T₂ temperature minus initial T₂ temperature), so that essentially cooling referred to a negative T₂ temperature change. An estimate of the vertical temperature gradient was also found based on the temperature difference T₁–T₂ at the end of the 2 min interval. The experiments for each configuration were repeated to verify the consistency of the results. The repeated measurements showed temperature variations of no more than 0.5 °C, indicating good repeatability.

In the configuration utilizing five thermocouples, only the highest wind speed of 9.04 m/s was examined. In this case, the wind tunnel was turned on once the catcher shaft temperature reached 39 °C ± 0.5 °C. Typically, a vertical thermal gradient was observed, with temperatures at the ceiling (sensors 1 and 2) consistently exceeding those at the floor (sensors 4 and 5). Thermal symmetry was maintained relative to the vertical centerline of the window for the 0° and 180° orientations, for which the maximum temperature deviation between symmetrical sensors (1 vs. 2 and 4 vs. 5) was 0.3 °C. In other orientations, this difference increased to a maximum of 3 °C. Regarding temporal stability, the system reached a steady state following a transient period of about 5 min.

3. Results

The results for the closed- and open-cavity cases are summarized in this section.

3.1. Closed Cavity

The results presented in Figure 5a show that, regardless of the wind speed magnitude, the greatest cooling effect when the cavity is closed at its back appears for a wind direction of 180° (α = 180°), measuring a 4 °C temperature drop in T₂. For this orientation, the temperature at the catcher inlet (T₁) was more than 5 °C higher than at the floor (as shown in Figure 5b), creating a thermal gradient that drives upward airflow. Additionally, because the catcher’s opening at α = 180° lies in the wake of the structure, it is exposed to lower static pressures (suction), which further enhances the internal airflow. These two mechanisms, namely the temperature-induced buoyant flow and the wind-induced suction, work together to maximize the cooling effect.

In contrast, the 0° orientation exhibits an inverse (negative) vertical temperature gradient, with the temperature at the top being more than 7.5 °C lower than at the floor level. Additionally, the pressure at the catcher’s opening is highest due to the formation of a stagnation point at the inlet. As a result, neither the thermal gradient nor the wind-induced pressure spatial difference contributes positively to internal airflow. Consequently, the cooling effect is reduced, with a maximum temperature drop in T₂ by only 1.7 °C at the highest tested wind speed (9.04 m/s). In fact, for wind speeds below 5.7 m/s, the floor-level temperature T₂ actually increases (positive temperature change, Figure 5a), indicating that the catcher heats the air rather than cooling it under this condition.

The worst performance is observed for the 90° orientation (α = 90°), where the system heats the air instead of cooling it. As shown in Figure 5a, the ground-level temperature increases with wind speed for this orientation. The vertical temperature gradient is positive up to a wind speed of 6.5 m/s, but reverses at higher speeds (Figure 5b). Overall, for the closed-cavity configuration, the best cooling performance is achieved at α = 180°, and it progressively deteriorates as the orientation shifts through α = 135°, α = 0°, and α = 45°, reaching its lowest efficiency at α = 90°.

Figure 6 illustrates the temporal temperature variations in a cavity cross-section (20 mm from the resistors towards the cavity exit), namely at location 1 (close to the cavity ceiling), at location 3 (section center), at location 4 (close to the floor), as well as at the catcher shaft, for two model orientations, 180° and 90°. For all cases, there is typically a vertical temperature gradient in the cross-section, with the higher temperatures close to the ceiling. Only for the 180° orientation, the temperature appears to be maximum at the central point (location 3). In the steady state, at 180°, the temperatures are smaller compared to their initial values (the air is cooled off), whereas at 90°, they are elevated (the air is heated), illustrating the better cooling performance at 180°. The catcher shaft temperature is reduced in all cases with time, being stabilized after a transient period of 4 to 5 min (see Figure 6).

Based on the temperatures of the five sensors, the mean temperature of the cross-section was estimated at t = 0 and at t = 6 min, the difference of which is shown in Figure 7a for each orientation. The maximum temperature reduction of 3.94 °C happens when α = 180°; a reduction of 2.35 °C happens when α = 135°; at α = 0° the reduction is 0.5 °C; at α = 45°, it increases by 0.6 °C; and finally at α = 90°, the temperature is increased by 3.4 °C. These temperature changes reflect the same performance behavior of the wind catcher as that based on Figure 5a, albeit with different temperature variations. The corresponding temperature difference between the catcher shaft and the cavity cross-section is negative for the model orientations α = 0°, 45°, and 90° and positive for the rest of orientations, as shown in Figure 5b.

3.2. Open Cavity

In contrast to the closed-cavity configuration, where in some cases the air is heated rather than cooled, the open-back-side case consistently results in cooling (negative change in T₂), regardless of wind speed or orientation angle.

As shown in Figure 8a, the maximum reduction in the floor-level temperature T₂ occurs at an orientation of α = 135°, reaching 9 °C. At the same time, the temperature at the upper level is 2 °C higher than at the floor level (Figure 8b). A plausible explanation for this significant temperature drop, compared to the closed-cavity case, is that at this particular orientation, air enters the cavity from the rear and exits through the top opening of the wind catcher, effectively cooling the ground floor. Additionally, the temperature gradient assists air movement, as temperature increases with elevation along the catcher shaft, enhancing natural convection in the direction of the airflow (from lower to higher level).

In contrast, under a 0° wind orientation, the cooling effect is limited, with floor-level temperature reduced by almost 2 °C. In this configuration, air enters directly through the catcher opening and removes the heat from the electric resistors located both near the catcher inlet and at the floor level. As a result, the heated air flows over the area where the temperature is measured close to the floor, diminishing the overall cooling effect. Regarding the vertical temperature gradient, in this case, it is favorable in the sense that the temperature increases along the direction of airflow, from 3 °C to 5 °C, depending on the wind speed. The cooling performance is altered as the angle changes from α = 135° to α = 180°, where the temperature drop is close to 6 °C, then to α = 45°, where the temperature drop is between 4 °C and 5 °C, then to α = 90° (temperature drop of 3 °C to 4 °C), and finally to α = 0° (temperature drop of 2 °C). The better cooling performance at α = 135° compared to α = 180° might be attributed to the higher suction pressures at the catcher’s opening, due to its close proximity to the wind tunnel’s vertical walls, causing a local acceleration of the flow outside the model. This issue was clarified by using the five thermocouples, showing that the best performance appears for the 180° instead of the 135° orientation. As illustrated in Figure 9a, at α = 180°, the mean temperature at the cross-section drops by 5.9 °C (like in Figure 8a), but at α = 135°, it drops by 4.97 °C. Specifically, at α = 180°, the temperatures of all five sensors are equal, whereas at α = 135°, there is a small temperature non-uniformity showing higher temperatures (0.5 °C to 1 °C) at the window right side (sensors 3, 2 and 5) due to the lower fluid speeds at this part of the window caused by local flow separation. For the other orientations, the temperature drops even less, with the smallest reduction at α = 0° of 2.5 °C. Regarding the vertical temperature gradient, this is positive at α = 90°, 135° and 180°, whereas it is negative at α = 0° and 45°, similarly with Figure 8b.

It is also interesting to note that for the open-cavity case, besides the α = 135° orientation, the wind speed’s change influence on T₂ is minor, in contrast to the closed-cavity case.

4. Discussion

The model examined in this study represents the simplest form of a wind catcher, featuring only a single opening and lacking internal partitions within the vertical shaft. However, wind catchers typically include multiple openings and various internal partitions, which facilitate airflow by allowing fresh air to enter through one port and exit through another. This detail plays a significant role in their thermal performance. Of course, one-sided wind catchers do exist, although this is not a rule (Montazeri [10]). In the current study, for the closed-cavity configuration, the optimal cooling effect was observed when the wind catcher opening was positioned in the wake of the structure. In this scenario, the lower local external pressure effectively drew out the interior air, enhancing ventilation mainly through a suction mechanism. Similar observation was performed in Montazeri & Azizian [8], in which the orientation of α = 180° was found to be working effectively, with the catcher functioning as a suction device. Natural convection, in this case, plays a smaller role. In fact, at α = 180°, in the closed-cavity case, the removed heat due to natural convection was estimated to be about 15% of the total heat load in the present study. More particularly, the removed heat by convection from the vertical walls of the wind catcher shaft, is given by the following formula:

$$\dot{Q}=hA(T_s - T_{\infty })$$

(2)

where the exposed area A = 0.038$\times$3 $\times$ 0.08 = 0.009 m², considering that the three heated walls of 0.038 m width are 0.08 m long (as the part of the catcher’s opening is excluded). The average temperature of the wall where the heaters were installed was $T_s$ = 90 °C and the mean temperature of the air $T_{\infty }$ = 27.35 °C (based on the top and bottom air temperatures) at α = 180°. The heat transfer coefficient was calculated using the formula for vertical plates for natural convection (Table 9.1, Equation (9–19), Cengel, and Ghajar [29]), assuming a minimal interaction between the boundary layers of the three walls:

$$Nu=0.59·{Ra}^{1/4}$$

(3)

where the Rayleigh number $Ra=Gr·Pr$ and the Grashof number is $Gr=g$β$\left(T_s-T_{\infty }\right)l^3/{\nu }^2$, where $l$ = 0.08 m (the length of each of the four sides of the shaft), β is the air thermal expansion coefficient, and $\nu$ is the air kinematic viscosity at the average temperature ($T_s+T_{\infty })/2$ = 58.67 °C. Based on the above, $Ra=1.9\times {10}^6$ and the heat transfer coefficient $h=$ 7.67 W/m²·K. Therefore, based on (2), the heat removed by natural convection is 4.38 W or 14.60% of the total heat load. It is noted that the heat transfer coefficient is weakly dependent on the model size as well as the variation in the temperature difference ($T_s-T_{\infty })$ according to (3). In computational and experimental work by Carreto-Hernandez et al. [32], where a spray humidification system was applied in a 3.85 m high wind tower, the Grashof number varied from 10⁹ to 10¹⁰ and the Nusselt number had values up to 200 (Figure 12 in [32]). The latter values, much higher compared to this work are mainly due to the small scale of the model. However, the heat transfer coefficient values in [32] are comparable with the present ones, not exceeding 11 W/m²·K (Figure 15 in [32]).

It is also important to highlight that the incoming air stream used in the experiments had a temperature of 19 °C. As reported in Chohan et al. [14], the optimal performance of wind catchers in the UAE occurs when ambient nighttime temperatures range between 16 °C and 24 °C, causing a maximum interior temperature reduction of 7 °C. In comparison, the current study demonstrated a maximum temperature reduction of 4 °C in the closed-cavity configuration and up to 6 °C in the open-cavity case.

It should be noted that in all fifty cases examined, the temperatures at the moment the wind tunnel was activated were T₁ = 42 °C and T₂ = 26 °C, with a maximum deviation of 0.5 °C. The higher value of T₁ compared to T₂ was due to the closer proximity of the thermocouple to the three resistors in the catcher’s shaft. The initial value of T₂ was selected to be higher than the temperature of the free stream. Although these initial temperatures were selected arbitrarily, they were kept constant across all tests to ensure meaningful comparisons. The conclusions in this study are drawn based on temperature measurements taken after a 2 min operation period of the wind tunnel. When this period was increased to 4 min, the trends remained the same. Furthermore, when the location of temperature measurement was changed to the middle of the cavity (from the bottom wall to mid-height) for the highest wind speed case (9.04 m/s), the trends remained the same. As it is shown in the computational work of Foroozesh et al. [16], Figure 3, the temperature field is quite uniform from the middle of the cavity till its exit, explaining why the change of the location did not alter the results. Moreover, the spatial temperature distribution was also evaluated for the highest examined free-stream velocity of 9 m/s by taking measurements at five points along a vertical cross-section of the cavity located 20 mm downstream of the resistors. In the open-cavity configuration, the temperature field remained quite uniform, with point-to-point variations not exceeding 2 °C. Conversely, the closed-cavity configuration exhibited temperature non-uniformities up to 5 °C, particularly at the 90° orientation. Despite these localized variations, the spatial mean temperature yielded conclusions regarding cooling performance consistent with single-sensor data. Moreover, it was clarified that for the open-cavity case, the 180° orientation corresponds to the maximum cooling performance, instead of the 135° one. The temperature non-uniformities have been predicted by several published works (Nejat et al. [33]; Sangdeh & Nasrollahi [34]; Shayegani et al. [35]), and future work aims to expand the number of temperature measurement points, enabling a more accurate assessment of the wind catcher’s thermal performance.

5. Conclusions

The cooling performance of a single-opening wind catcher was experimentally investigated in a wind tunnel using a wooden model. Electric resistors were employed to simulate heat loads at both the catcher shaft and a horizontal cavity connected to the catcher. The study focused on assessing the influence of wind speed and wind direction relative to the catcher opening on cooling performance. Wind speeds ranging from 4 m/s to 9 m/s were tested, along with five wind orientation angles with respect to the catcher’s opening: 0°, 45°, 90°, 135°, and 180°. Two cases were examined, namely one considering the back end of the cavity closed and the other open, respectively.

5.1. Closed Cavity

The greatest cooling effect within the cavity (maximum temperature drop of 4 °C) was observed when the wind direction was at 180°. In this case, the wind catcher opening was located in the wake of the vertical shaft, resulting in a low-pressure zone that, combined with a favorable vertical temperature gradient, enhanced the induced air flow and cooling performance. The poorest cooling performance was observed at the 90° wind orientation, at which the air within the cavity was heated rather than cooled at all tested wind speeds. For the remaining wind directions, the performance was intermediate, improving with increasing wind speed.

5.2. Open Cavity

When the cavity was open, the interior temperature consistently decreased, regardless of the wind speed or orientation angle. The maximum temperature reduction of 6 °C was observed at 180° wind direction, higher than in the closed-cavity case. This enhanced performance is attributed to forced convection, as air entered from the rear of the cavity at α = 180°. Additionally, a favorable vertical temperature gradient, with higher temperatures at the top of the catcher, contributed to the improved cooling effect. Regarding the other orientations, the temperature reduction was smaller, down to 2 °C.

6. Limitations

The wind catcher is assumed to operate in isolation, exposed to winds over flat terrain. As a result, the effects of neighboring buildings are not included in the analysis, which is an important limitation when considering use in urban environments. However, for densely populated areas, the operation of the wind catcher would be simulated with the closed-cavity case, as the worst-case scenario.

7. Future Work

Future work will focus on expanding the spatial resolution of thermal measurements and exploring multi-opening designs. Additionally, the influence of neighboring buildings will be considered in the estimation of the wind catcher’s efficiency. To further isolate and quantify wind-tunnel wall effects, future work will involve experiments with a smaller-scale model and blockage ratio, thereby minimizing tunnel interference and enabling a more accurate assessment of orientation-dependent performance under representative conditions.