Airflow Characteristics According to the Change in the Height and Porous Rate of Building Roofs for Efficient Installation of Small Wind Power Generators

Abstract

In this paper, the air flow characteristics and the impact of wind power generators were analyzed according to the porosity and height of the parapet installed in the rooftop layer. The wind speed at the top was decreasing as the parapet was installed. However, the wind speed reduction effect was decreasing as the porosity rate increased. In addition, the increase in porosity significantly reduced turbulence intensity and reduced it by up to 40% compared to no railing. In the case of parapets with sufficient porosity, the effect of reducing turbulence intensity was also increased as the height increased. Therefore, it was confirmed that sufficient parapet height and high porosity reduce the effect of reducing wind speed by parapets and significantly reducing the turbulence intensity, which can provide homogeneous wind speed during installation of wind power generators.

1. Introduction

Due to industrial development, the awareness of environmental problems is growing all over the world. Among them, the damage caused by global warming is rapidly increasing, and solutions are being sought around the world to solve it. As a result, paradigm shifts are taking place in the field of urban architecture. Green cities and architecture are in the spotlight all over the world to increase energy efficiency and ease environmental problems by utilizing renewable energy. Among the renewable energy used in green cities and architecture, wind energy is one of the essential factors because it has great economic potential and does not cause greenhouse effects.

Many studies have been conducted through wind tunnel experiments and computerized fluid dynamics on optimal location selection and on airflow characteristics around buildings to apply small wind power generation in urban areas. Richard F. Smith (2007) [1] studied the analysis of airflows and the efficiency of installed wind turbines at the Bahrain World Trade Center via CFD (Computational Fluid Dynamics). Roy Denoon (2008) [2] studied the efficient building form and layout during wind power generation through CFD and wind tunnel experiments. Lin Lu and Ka Yan Ip (2009) [3] conducted a study of the wind speed between two buildings and the wind speed according to the height and roof shape of the building through CFD, indicating a three to eightfold increase in wind density depending on the height and roof shape of the building. Khayrullina, A. [4] studied the efficiency of various wind turbines installed between buildings, and the location of wind turbines according to wind direction angles, turbulence, and the type of buildings. Ledo et al. [5] conducted a study on the air flow characteristics of the wind power generator installed on the roof of a building, indicating that a flat roof has a larger wind energy density than other roofs and remains constant. Dursun Ayhan and Safak Salglam [6] analyzed the efficiency of wind turbines installed in downtown buildings according to the geometry of buildings through CFD. The analysis showed that wind turbines increased up to eight times as high as buildings. Abohela et al. [7] conducted a study to select the optimal location for installing small wind power generators on roofs according to different roof shapes and wind angles, showing that the maximum acceleration of wind speed occurred on arched roofs, which can increase energy acquisition by up to 56.1%. In order to utilize wind energy in urban areas, Campbell et al. [8] proposed three methods for applying wind power systems:

(1)

Independent installation of wind power generators through site selection in urban areas;

(2)

Installation of wind power generators in existing buildings;

(3)

Integrate wind generators into buildings (Building Integrated Wind Power).

Buildings are more concentrated in downtown areas, so there is a great difficulty in selecting sites to install wind power generators independently. Building Integrated Wind Power (BIWP), which incorporates wind turbines into buildings, is economical and efficient without the need for support to position turbines to installation height. However, the large scale of wind power generation can cause noise and vibration. In recent years, micro wind power generation systems (Micro Wind Turbine) have mainly been applied. Small wind power generators are defined as having a performance of 2.5 kW or less [9], or a rotor diameter of 1.25 m or less [10].

Generally, the Reynold-Average Navier–Stokes (RANS) simulation is the most widely used method for modeling the flow turbulence due to the well-developed CFD best practice guidelines, but it has some inherent limitations for modeling complex flow and unsteady flow structures [11,12,13,14]. The direct numerical simulation (DNS) would be prohibitive for flow with high Reynolds numbers in present study. Thus, we choose the LES (Large Eddy Simuation) approach to solve the problem with affordable computational costs, which is also considered as the approach with the most potential by many researchers [11,14]. Generally, the LES research on the flow field around low-rise buildings mainly focuses on the case of an isolated building without snowdrift [15,16,17,18,19]. Some scholars have considered the influence of the surrounding buildings on the flow characteristics [20,21,22]. These numerical studies have been well verified by wind tunnel experiments or field measurements.

According to Ledo et al. [23], the installation of small wind power generators in urban buildings is subject to restrictions due to low wind speeds, high turbulence intensity, and recognition of noise generated by turbines. In addition, air flow characteristics around buildings vary depending on the shape or size of buildings, layout, spacing, wind direction, and surface classification [24,25,26]. Therefore, research on optimal installation locations with high wind speeds and low turbulence intensity is needed to increase the efficiency of wind turbines. There are various ways in which small wind farms are installed in buildings. In this paper, wind tunnel experiments and CFD analyses were conducted to analyze the effects of wind speed and turbulence on small wind power generators installed on rooftops due to changes in the height and porosity of mid- to low-rise buildings in the city.

As such, many studies have been conducted to apply wind power generators in downtown buildings and to check the efficiency of wind power generation for various roof types of buildings. However, a safety railing is installed on the rooftop of the actual building. If a railing is installed on the rooftop floor of a building in the city center, it will have different air flow characteristics from the top of the rooftop; thus, the analysis of air flow in the rooftop floor of a building with railings is insufficient. Therefore, in this paper, we want to analyze the wind tunnel experiment and computational fluid dynamics of the upper air flow characteristics according to the porosity and height of the parapet installed in the building for optimal efficiency when installing the actual wind power generator.

2. Experimental Specification of the Target Building

The basic type of the building to be interpreted is a cuboid building with no railing, and it is approximately 6 m high, or approximately two stories. Handrails installed on the upper part of the rooftop were carried out for cases where porosity was not present and for cases where porosity was present, and the porosity rate of the handrails was modelled to have 30%, 50%, and 80% porosity. The height of the railing was set at 0.1, 0.2, and 0.3 times the height of the building (${\mathrm{H}}_{\mathrm{b}}$). In this study, a total of 15 cases were analyzed depending on the change in the aspect ratio of buildings, the type of railing by type, and the height of the railing. Figure 1 shows the case definition, and

Table 1. Classification of cases.

Case	Porosity	$\mathbf{Height} \mathbf{of} \mathbf{Parapet} \left({\mathbf{H}}_{\mathbf{p}}\right)$
Case N	No parapet
Case 0–0.1	0	0.1 ${\mathrm{H}}_{\mathrm{p}}$
Case 30–0.1	30
Case 50–0.1	50
Case 80–0.1	80
Case 0–0.2	0	0.2 ${\mathrm{H}}_{\mathrm{p}}$
Case 30–0.2	30
Case 50–0.2	50
Case 80–0.2	80
Case 0–0.3	0	0.3 ${\mathrm{H}}_{\mathrm{p}}$
Case 30–0.3	30
Case 50–0.3	50
Case 80–0.3	80

shows the case classification according to the case definition.

3. Experimental and Numerical Validations

Computational fluid analysis (CFD) requires less time and cost than wind tunnel experiments, and numerical data can be obtained from anywhere in the three-dimensional space without repetitive tasks on the same object. It also has the advantage of being able to easily identify the characteristics of the airflow through the visualization of these numerical data. However, if an accurate Boundary Condition is not given in the computational fluid analysis, the results can be different from the actual ones. Therefore, the validity check of the computational fluid analysis was carried out prior to conducting the CFD for this study. The validity check measured the average wind speed and turbulence intensity on the roof of the building using a wind tunnel experiment device and analyzed the degree of correlation by obtaining the results and correlation coefficients of the computational fluid analysis. The correlation coefficient (${\mathsf{\gamma }}_{\mathrm{xy}}$) is defined by the following expression (1):

$${\mathsf{\gamma }}_{\mathrm{xy}}=\frac{{\sum }_{i=1}^n(x_i-\overline{x})\left(y_i-\overline{y}\right)}{\left(n-1\right)S_x{S}_y}$$

(1)

where $x_i, y_{i }$ = average value of x and y, respectively, and $S_x, S_y$ = standard deviation of x and y, respectively.

The results of the CFD and wind tunnel experiments were dimensionless by dividing the results of the average wind speed and turbulence intensity by the height of the building at the top of the rooftop by the average wind speed and turbulence intensity measured at each height. The dimensionless wind velocity and turbulence intensity ratios are defined by the following expressions (2) and (3):

$${\mathrm{V}}_{\mathrm{N}}=\frac{V_{Building}}{V_{Empty}}$$

(2)

$${\mathrm{I}}_{\mathrm{N}}=\frac{I_{Building}}{I_{Empty}}$$

(3)

where ${\mathrm{V}}_{\mathrm{N}}, {\mathrm{I}}_{\mathrm{N}}$ is wind speed ratio and turbulent intensity ratio, $V_{Building}, I_{Building}$ is wind speed ratio and turbulent intensity ratio by measurement height at the top of the rooftop, and $V_{Empty}, I_{Empty}$ is wind speed ratio and turbulent intensity ratio by measurement height in case of no building.

3.1. Wind Tunnel Experiment

The wind tunnel experiment was conducted in an open wind tunnel device with a measuring section of 2.0 m (width) × 1.7 m (height) × 12 m (length) held by the wind tunnel laboratory at Jeonbuk National University. Figure 2 shows the experimental wind direction angle. Wind pressure tests were conducted in seven directions, with wind test angles varying from 0° to 90° at intervals of 15°, centered on the pilot position. The length of the tube used in the wind pressure test was 120 cm, and the wind pressure signal was corrected using a resistor at a specific location in the tube. Figure 3 shows the effect of correction for the pressure transfer characteristics of the gain and phase angle used in the experiment. With regard to the gain, the blue signal indicates the resonance shape generated in the tube before the resistance tube was installed, whereas the red signal indicates the corrected signal after the resistance tube was installed. Each wind pressure measured in the wind pressure test is displayed as a dimensionless value.

Table 2. Wind tunnel experimental conditions.

Measurement Apparatus	Series 100 Cobra Probe
Surface roughness	D (α = 0.10)	B (α = 0.22)
Reference wind speed (${\mathrm{U}}_{\mathrm{ref}}$)	5 m/s
Wind angle	45°
Measurement frequency	150 Hz (60 s)
Case	80–0.2

shows the experimental conditions for wind tunnel experiments. The experimental building is Case 80–0.2 and the experimental scale is 1/30. Figure 4 shows the distribution of wind speed and turbulence intensity by height reproduced within the wind tunnel. The surface roughness used in the wind tunnel experiment was conducted mainly on the coast (surface roughness D) and urban areas (surface roughness B). Figure 5 shows the position and height of measurement at the top of the rooftop. A total of 25 measurement points were allocated evenly on one side of the building to five measurement locations on the upper part of the building. In addition, the measurement height has been increased a total of six times by 0.1 times the height of the basic building (${\mathrm{H}}_{\mathrm{b}}$), with the measurement height starting at 1.1 H and the final measurement height being 1.6 ${\mathrm{H}}_{\mathrm{b}}$.

3.2. CFD Model

This investigation used a geometric model that combines elements of two mesoscales. There is also a microscale model. Geometric models are the same as the microscale models that include explicitly configured building details, but the compute domain extends as follows: The local field domain for the weather station several kilometers away is much larger. Used by microscale models, this includes a small area of computation around it. Thus, we call this a full-scale model that differs from the micro-scale model only in the computational domain size. This work is based on the recommendations from an overview of CFD use in simulations. The contents of the outdoor environment [27], “CFD Best Practice Guidelines” by Blocken et al. [28], and the Architectural Society Flow Simulation of Japan (AIJ) Pedestrian Environment CFD Practical Use Guidelines Building [29]. These simulations were carried out using commercial CFD (program, ANSYS Fluent 16.1) [30]. The CFD model used in this program solves the Reynolds-averaged Navier–Stokes (RANS) equations with the renormalization group (RNG) $k-\epsilon$ turbulence model. The model performed well in simulations of urban wind flows [23,31]. In the RNG $k-\epsilon$ model, the turbulent kinetic energy k and its transportation equation dissipation rate $\mathsf{\epsilon }$ are:

$$\frac{\partial \left(\rho k\right)}{\partial t}+\frac{\partial \left(\rho k{u}_i\right)}{\partial x_i}=\frac{\partial }{\partial x_j}\left[(\nu +\frac{{\nu }_t}{{\sigma }_k})\frac{\partial \left(k\right)}{\partial x_j}\right]+G-\epsilon$$

(4)

$$\frac{\partial \left(\rho \epsilon \right)}{\partial t}+\frac{\partial \left(\rho \epsilon u_i\right)}{\partial x_i}=\frac{\partial }{\partial x_j}\left[(\nu +\frac{{\nu }_t}{{\sigma }_{\epsilon }})\frac{\partial \left(\epsilon \right)}{\partial x_j}\right]+C_{\epsilon 1}\frac{\epsilon }{\kappa }G-C_{\epsilon 2}\frac{{\epsilon }^2}{\kappa }-C_{\mu }{\eta }^3\frac{1-\eta /{\eta }_0}{1+\beta {\eta }^3}\frac{{\epsilon }^2}{\kappa }$$

(5)

where $\rho$ is air density (kg/m³); t is time (s); $u_i$ and $u_j$ are the Reynolds time-averaged velocity component in the $x_i$ and $u_{j }$ (i, j = 1, 2, 3) directions, respectively; $\nu$ is the dynamic viscosity of air (m²/s); ${\nu }_t=C_{\mu }\frac{{\kappa }^2}{\epsilon }$, the turbulence kinematic viscosity (m²/s); ${\sigma }_k=$ 0.7194 is the turbulence effective Prandtl number for k; G is the source term; and ${\sigma }_{\epsilon }=0.7194$ is the turbulence effective Prandtl number for

$$\epsilon ,{ C}_{\epsilon 1}=1.42, C_{\epsilon 2}=1.68, C_{\mu }=0.085,\eta ={\left(2E_{ij}·E_{ij}\right)}^{1/2}\frac{\kappa }{\epsilon }=\frac{1}{2}\left(\frac{\partial \left(u_i\right)}{\partial x_j}+\frac{\partial \left(u_j\right)}{\partial x_{i }}\right), \beta =0.012, and {\eta }_0=4.38.$$

(6)

The analysis was conducted by modeling it on the same scale as the wind tunnel experiment 1/30 to validate its validity with the wind tunnel experiment.

Table 3. Input condition for computational fluid analysis.

Solution Method	Second Order
Residual	No Criteria
Turbulence model	Realizable $k-\epsilon$
Reference Velocity ($V_{ref}$)	5 m/s
Iteration	1000
Boundary condition	Inlet	User-Defined Function
	Outlet	Pressure outlet
	Top and side	Specified shear wall
	Bottom	No-slip wall with standard wall functions

shows the conditions for interpretation of computational fluid analysis. The average wind velocity, turbulence energy, and turbulence dissipation rates of surface roughness D and B reproduced in the wind tunnel were entered through the User-Defined Function (UDF). Figure 6 shows the vertical distribution of the average wind velocity, turbulent energy, and turbulent dissipation rate of the incoming airflow.

Grid Sensitivity Verification

The lattice sensitivity test was conducted to determine the size and shape of the most appropriate grid, considering the time required for computer performance and analysis during computer fluid analysis and the accuracy of the analysis results. A total of three types of lattice sensitivity tests were conducted, and the highest correlation type of lattice size and shape were reflected in the computational fluid analysis of this study, correlating with 45 degrees of wind angle of Case N as shown in

Table 4. The size, number, and shape of the grid by type.

Type	1	2	3
the perimeter of a building	1 cm	2 cm	2 cm
External flow	4 cm	4 cm	3 cm
Number of nodes	5,079,016	1,175,377	884,195
Number of lattices	4,974,356	1,135,134	4,659,800
Grid form	Hexa	Hexa	Tetra

, where there is no handrail.

Table 5. Comparison of correlation coefficients by case.

Type	1	2	3
Wind speed ratio	0.889	0.942	0.405
Turbulent intensity ratio	0.913	0.892	0.663

shows the results of the correlation coefficient by case, and Figure 7 shows the comparison graph of the correlation coefficient of wind speed ratio and turbulence intensity ratio by case. Type 1 and 2 exhibit similar accuracy in correlation. However, it takes about 2 to 3 times more time to interpret Type 1 compared to Type 2. Therefore, in this study, Type 2 was used to determine the shape and size of the grid and to conduct an interpretation. Figure 8 shows a case-by-case correlation graph.

3.3. Verification Results

Figure 9 shows the correlation between computational fluid analysis and wind tunnel test results. At surface roughness D, wind speed ratio was 0.884, turbulence intensity ratio was 0.848, wind speed ratio was 0.945, and turbulence intensity ratio was 0.801, indicating a high positive correlation in both surface roughness D and B. Therefore, it is deemed that the computational fluid analysis in this study will be applicable to the analysis of airflow characteristics in the upper part of the rooftop.

4. Results and Discussions

A computational fluid analysis was conducted to determine the optimal location of wind power generators by analyzing the air flow characteristics of the upper air flow rate and height of the parapet installed on the upper roof of the actual building. Although wind speed contributes greatly to the performance of wind power generators, the impact on the fatigue life of wind power generators should be minimized by providing homogeneous wind quality by identifying locations where turbulence is less affected. To minimize the effects of turbulence, Islam Abohela [7] conducted a study on the location of rooftop wind turbines according to the roof types of various buildings. When wind turbines are installed, turbulence intensity is greater than 1.3 H. In addition, according to the Encraft Warwick Wind Trials Project [13], the lowest position of a wind generator rotor is proposed to be at least 30% of the minimum building height at the top of the roof, and the WINEUR [14] report suggests 35% to 50% of the building height. Therefore, in this study, less than 1.3 H, which produces significant turbulence intensity, was named Turbulence Area, and wind speed ratio in the turbulent area was not considered.

4.1. Selection of Wind Angle

In order to find the most advantageous wind angle when installing wind power generators, the analysis was conducted on wind direction angles of 0 and 45 degrees. Since the performance of the wind generator is greatly contributed to by the average wind speed, the surface roughness was interpreted for the surface roughness D, where the average wind speed is dominant. The conditions used in the analysis are shown in Table 3.

Table 6. Maximum wind speed ratio and turbulent intensity ratio according to wind direction and location of occurrence.

Type of Roof	Flat Roof
Wind angle	0 Deg		45 Deg
Maximum wind speed ratio/Turbulent intensity ratio	1.136	1.969	1.132	1.350
Location	(2,3)		(5,5)
Height	1.3 ${\mathrm{H}}_{\mathrm{b}}$		1.3 ${\mathrm{H}}_{\mathrm{b}}$

shows the maximum wind speed ratio and turbulence intensity ratio and location of the roof layer of the building according to the wind direction. At 0 degrees wind direction angle, the maximum wind velocity ratio is 1.3 ${\mathrm{H}}_{\mathrm{b}}$ at the top of the turbulence zone. At 45 degrees wind direction angle, the maximum wind velocity ratio is at (55) of 1.3 ${\mathrm{H}}_{\mathrm{b}}$ at the top of the turbulence zone. In addition, the maximum wind speed ratio is 0.37% smaller than the wind direction angle of 0 degrees. The maximum wind speed ratio according to changing angle of wind direction was not significantly changed within 1%. However, in the case of turbulent intensity ratio, wind direction angle of 45 degrees was 31% smaller than that of 0 degrees. Therefore, the difference in the maximum wind speed ratio is not significant, and the wind angle of 45 degrees, where the maximum wind speed ratio occurs, is more advantageous when installing wind power generators than a wind speed of 0 degrees. Figure 10 shows the maximum wind velocity ratio and turbulence intensity ratio at the top of the rooftop according to the change in height of the wind direction angle.

4.2. Parapet Impact Analysis

In order to apply small wind power generation from the rooftop of the building, the air flow analysis was conducted on the upper part of the rooftop. In fact, railings are installed on the rooftop floor of a flat roof building in accordance with the Enforcement Decree of the Building Act for safety. The installation of these handrails has a significant effect on airflow at the top of the rooftop. A computational fluid analysis was performed to analyze the air flow characteristics of the upper roof layer according to the porous rate of the handrail and the height of the handrail.

The impact analysis of the parapet installed on the upper part of the rooftop was conducted on the same location and height where the maximum wind speed occurs on the flat roof without the railing installed.

Table 7. Location and height of maximum wind speed ratio of Case N.

Surface Roughness	$\mathbf{D} (\mathsf{\alpha } = 0.10)$	$\mathbf{B} (\mathsf{\alpha } = 0.22)$
Location	5,5
Height	1.3 H	1.4 H
Maximum wind speed ratio	1.132	1.096
Turbulent intensity ratio	1.350	1.097

shows the location and height of the maximum wind speed ratio of Case N and the maximum wind speed ratio and turbulence intensity ratio.

Table 8. Wind speed ratio and turbulent intensity ratio for cases with parapet.

Surface Roughness	D			B
Case	$V_N$	$I_N$	Height	$V_N$	$I_N$	Height
Case N	1.132	1.350	1.3 H	1.096	1.097	1.3 H
Case 0–0.1	1.026	1.648		1.059	1.211
Case 30–0.1	1.03	1.639		1.065	1.159
Case 50–0.1	1.069	1.342		1.068	1.111
Case 80–0.1	1.085	1.335		1.075	1.083
Case 0–0.2	0.795	2.715		1.010	1.310
Case 30–0.2	0.64	3.868		0.997	1.280
Case 50–0.2	0.909	1.447		1.030	1.066
Case 80–0.2	1.076	0.904		1.065	0.969
Case 0–0.3	0.772	2.754	1.4 H	0.836	1.773
Case 30–0.3	0.798	2.384		0.714	2.139
Case 50–0.3	0.871	1.296		0.869	1.174
Case 80–0.3	1.036	0.764		1.018	0.792

shows the wind speed ratio and turbulence intensity ratio of the case with the parapet installed at the same height and location of Case N. The height of the analysis of the case with a parapet height of 0.3 ${\mathrm{H}}_{\mathrm{b}}$ at the surface roughness D was 1.4 H because the railing is located at the height of Case N.

4.2.1. Parapet Height

Figure 11 and Figure 12 show the wind speed ratio and turbulence intensity ratio of porous stars according to the height of the parapet by surface roughness. Regardless of the height and porosity of the parapet, wind speed ratio is lower than that of no railing. As the height of the parapet increases, the wind speed ratio decreases, and the turbulence intensity ratio tends to increase in cases with low porosity. However, if the porosity rate is 80%, the wind speed reduction effect by the parapet is reduced, and the turbulence intensity ratio decreases as the height of the parapet increases, and is lower than Case N. The effect of reducing turbulence intensity by parapet is significant from 0.2 H to 0.3 H, resulting in a 33% and 43% decrease in surface roughness D, respectively, and an 11% and 27% decrease in surface roughness B, respectively. Therefore, as the height of the parapet with sufficient porosity increases, the effect of reducing turbulence intensity is also increased.

4.2.2. Parapet Porosity

Figure 13 and Figure 14 show the wind speed and turbulence intensity ratios by parapet height according to changes in porosity by surface roughness. As the porous rate of handrails increases, wind speed ratio increases, and turbulence intensity ratio decreases. If the porosity rate is low at 0% and 30%, the turbulence intensity ratio is higher than Case N, regardless of the height of the handrail, and about three times higher in Case 30–0.2. However, if the porosity rate is high at 80%, the turbulence intensity ratio is lower than that of no railing, and it decreases by up to 43%. Therefore, the increase in the parapet porosity results in a significant reduction in turbulence intensity.

4.2.3. Airflow Distribution inside Parapet

The air flow distribution was shown to determine the effect of air flow inside the roof top layer parapet. The target case has a height of 0.3 H and was shown for 30%, 50%, and 80% to see the effect of the porosity. The air distribution was represented by horizontal and vertical distribution, and the horizontal distribution was represented at 0.15 H, one-half the height of the parapet, while the vertical distribution was represented in the middle of the building, parallel to the wind direction. Figure 15 shows the vertical distribution representation area. Figure 16 and Figure 17 show the airflow flow diagram inside the parapet of surface roughness D and B. As the porosity increases, both D and B exhibit similar airflow patterns. Case 30–0.3 shows that vortex occurs widely in the middle of the rooftop. The increase in porosity shows that in Case 50–0.3 the wind velocity of airflow through the parapet increases, dividing the vortex around the middle, and in Case 80–0.3 the range of the vortex becomes narrow and dissipates.

5. Conclusions

The following conclusions were obtained from CFD analysis to analyze the effects of wind speed and turbulence of small wind power generators installed on the roof due to changes in the height and porosity of buildings in the city.

• The parapet without porosity installed on the roof was decreasing wind speed around the roof. However, as the porosity of the parapet increases, the rate of wind speed reduction has been shown to decrease. Furthermore, the increase in porosity showed that turbulence intensity decreased by up to 43% and 27% in surface roughness D (coastal area) and surface roughness B (city area), respectively.
• Wind speed ratio decreases as the height of the parapet increases, but turbulence intensity varies depending on the porosity. If the porosity of the parapet is low, the turbulence intensity ratio increases, but if the porosity is high, it decreases. The effect of reducing turbulence intensity by parapet is significant from 0.2 H and 0.3 H, resulting in a 33% and 43% decrease in surface roughness D, respectively, and an 11% and 27% decrease in surface roughness B, respectively. Thus, it can be seen that for parapets with sufficient porosity, the effect of reducing turbulence intensity increases as the height increases.
• Increasing the height and porosity of the rooftop parapet will reduce wind speed and decrease the size of turbulence intensity, which will increase the operation efficiency of the small wind power generator installed on the roof of the building.