Ice Distribution Characteristics on the DU25 and NACA63-215 Airfoil Surfaces of Wind Turbines as Affected by Ambient Temperature and Angle of Attack

Abstract

Icing on wind turbines reduces power generation efficiency and leads to safety issues. Consequently, in this paper, ice distribution characteristics on DU25 and NACA63-215 airfoils at ambient temperatures and angles of attack are explored VIA numerical simulation. The findings indicate that when the ambient temperature changes in the range of 248–268 K, the ice distribution range on the upper surface of the DU25 airfoil (0–3.07 mm) is wider than that of the NACA63-215 airfoil (0–1.91 mm), while the ice distribution range on the lower surface of the DU25 airfoil (0–12.13 mm) is narrower than that of the NACA63-215 airfoil (0–15.18 mm) due to the discrepancy in droplet collection efficiency and droplet freezing rate caused by airfoil structure and ambient temperature, respectively. At an angle of attack of 0°, the ice distribution range on the upper surface of the DU25 airfoil is almost the same as that of the NACA63-215 airfoil. At an angle of attack of 8°, the ice distribution range on the upper surface of the DU25 airfoil (0–1.05 mm) is broader than that of the NACA63-215 airfoil (0–0.675 mm), whereas the ice distribution range on the lower surface of the DU25 airfoil (0–17 mm) is narrower than that of the NACA63-215 airfoil (0–20 mm) due to the discrepancy in droplet collection efficiency caused by droplet flow trajectory. The angle of attack has a much greater effect on the peak ice thickness than ambient temperature. This study will provide guidance for the anti-icing coating design of wind turbine blades.

1. Introduction

Wind energy is the most promising renewable energy, mainly used in the field of wind power generation [1]. However, due to the frequent occurrence of extreme weather, ice formation occurs on the surface of wind turbine blades, which causes serious problems, including a decrease in power generation efficiency and a reduction in structural strength and lifetime [2,3,4]. Even worse, owing to centrifugal force, ice blocks on the surface of the blades may be shed, posing a threat to the safety of people in the surrounding area [5]. Therefore, exploring the ice distribution characteristics on blade surfaces is crucial for the anti-icing coating design of wind turbine blades.

The common methods for investigating the icing characteristics of wind turbine blades consist of icing wind tunnel experiments and numerical calculations. Experimentally, Li et al. [6] explored ice distribution characteristics on vertical-axis wind turbine blades under rotating conditions and found that non-uniform ice accumulation covered the entire blade, leading to variation in the blade airfoil. Shu et al. [7] reported that ice thickness increased sharply from the root to the middle of the wind turbine blade and then slowed down from the middle to the tip of the wind turbine blade. Gao et al. [8,9] examined the aerodynamic performance of iced blade airfoil under ice conditions and found a functional relationship between aerodynamic performance and angle of attack. Jin et al. [10,11,12] compared the ice distribution characteristics on different airfoil surfaces and analyzed the aerodynamic performance of different iced blade airfoils for a wind turbine. It was found that the aerodynamic performance of the iced S832 airfoil exhibited a more significant decrease than that of the iced S826 airfoil. Hu et al. [13] proposed a novel method for detecting ice thickness on the surface of wind turbine blades and reported a measurement error of 2.62%–6.0%. Sundaresan et al. [14] analyzed the effects of angle of attack and Reynolds number on the aerodynamic performance of wind turbine blades, and it was found that the lift coefficient showed less variation for iced airfoil at the stall angle of attack under rime ice conditions. Han et al. [15] investigated the effect of ice accumulation on the mechanical properties of wind turbine blades; it was found that there was a positive correlation between stress–strain and rotational speed.

To summarize, owing to experimental size limitations, scaled-down models of most wind turbine blades were employed in the experiments. To make the research more meaningful, it is necessary to establish similarity between scaled models and full-scale models. However, due to the unclear icing physical mechanism on wind turbine blades during the experimental process, it is difficult to establish a similarity between scaled models and full-scale models.

Owing to the fact that numerical calculations can reveal the icing physical mechanism of wind turbine blades, many scholars use numerical calculation methods to investigate ice distribution characteristics on the blade surfaces of wind turbines. Ibrahim et al. [16] employed numerical simulation to investigate the influence of blade design on icing characteristics under ice conditions. Wang et al. [17] proposed a numerical calculation method to examine the water film distribution and ice shapes on a NACA0012 airfoil under environmental conditions; it was found that liquid water content had a significant effect on horn ice shapes. Sheidani et al. [18] presented a numerical calculation approach to reveal the mechanism of the effect of ice accretion on the wake characteristics of NACA0021 airfoils; it was found that ice accretion led to power coefficient reduction due to wake vortex separation. Lu et al. [19] employed a numerical method to examine the influence of the angle of attack on the ice shapes of a NACA0012 airfoil, it was found that the mass of ice accretion was large at low angles of attack (4° and 8°). Xu et al. [20,21] developed an icing model to reveal the mechanism of the effect of icing parameters on the icing characteristics of NACA0012 airfoils; it was found that ice shapes were influenced by droplet collection efficiency and freezing rate under different icing conditions. Wang et al. [22] employed a numerical method to analyze the effect of the angle of attack on the ice distribution characteristics of a DU97 airfoil; it was found that the peak ice thickness at the trailing edge increased with the increase in the angle of attack.

From the literature review on numerical simulation, it is clear that most researchers focus on the physical mechanism of the ice distribution characteristics of symmetrical airfoils. However, there are some limitations in the above studies, which include:

(1)

There are few studies on the physical mechanism of the ice distribution characteristics of asymmetric airfoil surfaces;

(2)

The effect of ambient temperature and angles of attack on the ice distribution mechanism of the DU25 and NACA63-215 airfoil surfaces has not been reported.

Therefore, the main objectives of this study are as follows: (1) the numerical calculation method is employed to reveal the physical mechanism of the ice distribution characteristics of asymmetric airfoil surfaces, including DU25 and NACA63-215 airfoils, and (2) the effects of ambient temperatures and angles of attack on the icing distribution mechanism are explored.

2. Model

2.1. Mathematical Model

2.1.1. Airflow and Temperature Fields Model outside the Wind Turbine Blade

The prerequisite for exploring icing distribution characteristics on the wind turbine blade surface is to obtain airflow and temperature fields outside the wind turbine blade, which can be calculated according to mass, momentum, and energy conservation equations, which are written as follows [20,23]:

The mass equation is written as

$${\displaystyle \frac{\partial {\rho }_1}{\partial t}}+\nabla \cdot \left({\rho }_1\overrightarrow{V_1}\right)=0$$

(1)

The momentum equation is written as

$${\displaystyle \frac{\partial \left({\rho }_1\overrightarrow{V_1}\right)}{\partial t}}+\nabla \cdot \left({\rho }_1\overrightarrow{V_1}\overrightarrow{V_1}\right)={\sigma }^{ij}+{\rho }_1\overrightarrow{g}$$

(2)

$${\sigma }^{ij}=-{\sigma }^{ij}{p}_1+{\tau }^{ij}$$

(3)

The energy equation is written as

$${\displaystyle \frac{\partial \left({\rho }_1{E}_1\right)}{\partial t}}+\nabla \cdot \left({\rho }_1\overrightarrow{V_1}{H}_1\right)=\nabla \cdot \left[k_1\left(\nabla T_1\right)+V_i{\tau }^{ij}\right]+{\rho }_1\overrightarrow{V_1}\overrightarrow{g}$$

(4)

2.1.2. Droplet Flow Model

After determining the multi-physical fields outside the wind turbine blade, the next step is to calculate the flow characteristics of water droplets outside the wind turbine blade. Therefore, a droplet flow model was established according to Euler’s formula to address the droplet distribution on the wind turbine blade considering the droplet volume fraction, which is represented as follows [24]:

$${\displaystyle \frac{\partial \alpha }{\partial t}}+\nabla \cdot \left(\alpha \overrightarrow{V_2}\right)=0$$

(5)

$${\displaystyle \frac{\partial \left(\alpha \overrightarrow{V_2}\right)}{\partial t}}+\nabla \cdot \left(\alpha \overrightarrow{V_2}\overrightarrow{V_2}\right)={\displaystyle \frac{C_{\mathrm{D}}{\mathrm{Re}}_2}{24K}}\alpha \left(\overrightarrow{V_1}-\overrightarrow{V_2}\right)+\alpha \left(1-{\displaystyle \frac{{\rho }_1}{{\rho }_2}}\right){\displaystyle \frac{1}{{\mathrm{Fr}}^2}}$$

(6)

The dimensionless number in Equation (6) includes the Reynolds number, inertia parameter, and Froude number, which are expressed as follows:

$${\mathrm{Re}}_2={\displaystyle \frac{{\rho }_{1}d{V}_{1,\infty }\Vert \overrightarrow{V_1}-\overrightarrow{V_2}\Vert }{{\mu }_1}},K={\displaystyle \frac{{\rho }_2{d}^2{V}_{1,\infty }}{18L_{\infty }{\mu }_1}},\mathrm{Fr}={\displaystyle \frac{\Vert V_{1,\infty }\Vert }{\sqrt{L_{\infty }{g}_{\infty }}}}$$

(7)

The drag coefficient of water droplets in Equation (6) is based on an empirical correlation for flow around spherical droplets [25], which is given by

$$C_{\mathrm{D}}=\left({\displaystyle \frac{24}{{\mathrm{Re}}_2}}\right)\left(1+0.15{\mathrm{Re}}_2^{0.687}\right),{\mathrm{Re}}_2\le 1300$$

(8)

$$C_{\mathrm{D}}=0.4,{\mathrm{Re}}_2\ge 1300$$

(9)

2.1.3. Water Film and Ice Accretion Model

After determining the droplet flow characteristics outside the wind turbine blade, the next step is to calculate heat and mass transfer for water film and ice accretion. The water film velocity considering surface roughness is calculated by [26,27,28,29]

$$\overrightarrow{V_f}={\displaystyle \frac{y}{{\mu }_f}}{\overrightarrow{\tau }}_{1,\mathrm{wall}}$$

(10)

$${\tau }_{1,\mathrm{wall}}={\displaystyle \frac{1}{2}}{C}_{\mathrm{f}}{\rho }_1{V}_{\infty }^2$$

(11)

$$C_{\mathrm{f}}={\left[3.476-0.707\ln \left(k_{\mathrm{s}}/\mathrm{s}\right)\right]}^{-2.46}$$

(12)

The mass equation includes mass transfer due to droplet impact, water evaporation, and ice accretion, which is written as [20]

$${\rho }_{\mathrm{f}}\left[{\displaystyle \frac{\partial h_{\mathrm{f}}}{\partial t}}+\nabla \cdot \left(\overrightarrow{V_{\mathrm{f}}}{h}_{\mathrm{f}}\right)\right]=V_{\infty }\beta LWC-{\dot{m}}_{\mathrm{evap}}-{\dot{m}}_{\mathrm{ice}}$$

(13)

The energy equation includes heat transfer, phase change, ice accretion, convective heat flow, and radiant heat flow caused by water droplet impact, which is calculated by [20]

$$\begin{array}{l}{\rho }_{\mathrm{f}}\left[{\displaystyle \frac{\partial h_{\mathrm{f}}{c}_{\mathrm{f}}\tilde{T}}{\partial t}}+\nabla \cdot \left(\overrightarrow{V_{\mathrm{f}}}{h}_{\mathrm{f}}{c}_{\mathrm{f}}\tilde{T}\right)\right]=\left(c_{\mathrm{f}}{\tilde{T}}_{2,\infty }+{\displaystyle \frac{1}{2}}{\overrightarrow{V}}_2^2\right)V_{\infty }\beta LWC\\ -0.5{\dot{m}}_{\mathrm{evap}}\left(L_{\mathrm{evap}}-L_{\mathrm{sub}1}\right)+{\dot{m}}_{\mathrm{ice}}\left(L_{\mathrm{fus}}-c_{\mathrm{ice}}\tilde{T}\right)+\sigma \epsilon \left(T_{\infty }^4-T_{\mathrm{f}}^4\right)+{\dot{Q}}_{\mathrm{h}}\end{array}$$

(14)

2.2. Computational Domain and Mesh

The computational domain of DU25 and NACA63-215 blade airfoils commonly used for wind turbine blades are shown in Figure 1. The chord length (C) and span length of the DU25 and NACA63-215 airfoils are 0.1 m and 0.02 m, respectively. The computational domain is meshed by structured elements, and the normal distance of the first layer grid is 0.01 mm, as shown in Figure 2. The y+ of the grid is controlled to be around or below 1. The total number of grids in the flow field is 90945, which ensures high computational accuracy and improves computational efficiency.

2.3. Boundary Conditions and Solution Method

The far-field boundary conditions and wall boundary conditions were employed on the external flow field and solid surface of a wind turbine blade airfoil for numerical simulations, respectively. The icing conditions of the wind turbine are reported in ref. [30], including the ambient temperature of 233–323 K, the wind speed of 3–25 m/s, the LWC of 0–5 g/m³, and the MVD of 10–5000 μm. In this study, the icing conditions are given as follows: ambient temperature of 248–268 K, wind speed of 10 m/s, LWC of 1 g/m³, MVD of 20 μm, angle of attack (0°, 4° and 8°), and icing time of 30 min. The specific parameters are presented in

Table 1. Boundary conditions.

Types	Values
Far-field	Pressure: 101,325 Pa
	Temperature: 248–268 K
	LWC: 1 g/m3
	MVD: 20 μm
	Speed: 10 m/s
Wall	No slip

. The detailed solution method is reported in ref. [20].

2.4. Model Validation

The numerical simulation results were compared with experimental data in the NASA Icing Research Tunnel [31], and the icing conditions are given as follows: ambient temperature of 267 K, air speed of 102 m/s, LWC of 0.55 g/m³, MVD of 20 μm, icing time of 420 s, and angle of attack of 4°. It can be observed from Figure 3 that the simulated ice distribution is in good agreement with the experimental data. The error of the vast majority of points was within 2%. As shown in Figure 3, most of the droplet diameters are very close to the median volume diameters, which leads to the ice thickness distribution caused by water film solidification in the experiment being close to that in the numerical simulation. Notably, it was found that the ice horns can appear on the leading edge of the blade airfoil. In the experiment, the droplet size was different. On the other hand, in the numerical simulation, the different droplet sizes are equivalent to the medium volume diameter of the droplets. The experimental results indicate that different droplet sizes can affect the water film distribution on the airfoil surface, freezing into the thickest ice accretion at the location where the droplet size is largest under icing conditions due to the variation in heat and mass transfer. The distribution characteristics of ice accretion on the airfoil surface caused by water film solidification are described using the medium volume diameter of droplets in the numerical simulation. Therefore, this explains why the maximum error occurs at the horn ice.

3. Results and Discussion

In this section, we describe icing simulations performed to analyze the ice distribution characteristics on DU25 and NACA63-215 airfoil surfaces at various ambient temperatures and angles of attack.

3.1. Effect of Ambient Temperature

Figure 4, Figure 5 and Figure 6 depict the ice thickness distribution on DU25 and NACA63-215 airfoil surfaces at different ambient temperatures (248–268 K) for an angle of attack of 4°. For the upper surface, when the ambient temperature increases from 248 K to 268 K, the peak ice thickness on the leading edge of the DU25 and NACA63-215 airfoils decreases from 5.32 mm to 5.21 mm (2.1% reduction) and from 5.39 mm to 5.22 mm (3.2% reduction), respectively. For the lower surface, as the ambient temperature rises from 248 K to 268 K, the peak ice thickness on the leading edge of the DU25 and NACA63-215 airfoils drops from 6.21 mm to 5.91 mm (4.8% reduction) and from 6.52 mm to 6.12 mm (6.1% reduction), respectively. The peak ice thickness on the upper surface of the DU25 airfoil is almost the same as that of the NACA63-215 airfoil, while the peak ice thickness on the lower surface of the DU25 airfoil is smaller than that of the NACA63-215 airfoil. On the lower surface, it is noteworthy that the peak ice thickness on the trailing edge of the NACA63-215 airfoil (0.092 mm) is greater than that of the DU25 airfoil (0.077 mm). On the other hand, for the leading edge of the airfoils, the ice distribution range on the upper surface of the DU25 airfoil (0–3.07 mm) is wider than that of the NACA63-215 airfoil (0–1.91 mm), while the ice distribution range on the lower surface of the DU25 airfoil (0–12.13 mm) is narrower than that of the NACA63-215 airfoil (0–15.18 mm). Owing to the discrepancy in the geometric structure of the DU25 and NACA63-215 airfoils, the droplet flow trajectory outside various airfoils can exhibit significant differences, leading to a discrepancy in droplet collection efficiency distribution on DU25 and NACA63-215 airfoil surfaces. The ice accretion distribution on different airfoil surfaces caused by droplet freezing exhibits significant differences under icing conditions. For the trailing edge of airfoils, the ice distribution range of the DU25 airfoil (80.55–99.36 mm) is broader than that of the NACA63-215 airfoil (85.03–99.36 mm). This is ascribed to the fact that different airfoils can result in variation in droplet collection efficiency, freezing into non-uniform ice accretion under icing conditions.

3.2. Effect of Angle of Attack

The icing characteristics on the airfoil surface ARE not only affected by the ambient temperature but also by the angle of attack. The premise of exploring icing characteristics is to understand the droplet collection efficiency on various airfoil surfaces at different angles of attack. The droplet collection efficiency on DU25 and NACA63-215 airfoil surfaces at different angles of attack (0–8°) is displayed in Figure 7. It is obvious that the peak droplet collection efficiency on the surface of the NACA63-215 airfoil (0.72) is higher than that of the DU25 airfoil (0.59). Owing to the discrepancy in the geometric shapes of airfoils, droplet flow trajectories outside various airfoils can exhibit significant differences, which leads to a discrepancy in droplet collection distribution on airfoil surfaces. At an angle of attack of 0°, the peak droplet collection efficiency appears at the stagnation point. With an increase in the angle of attack, the peak droplet collection efficiency moves toward the lower surface of airfoils. Significantly, the peak droplet collection efficiency on the DU25 airfoil exhibits a larger offset than that of the NACA63-215 airfoil. This is significantly different from symmetric airfoils [20]. On the other hand, the distribution range of droplet collection efficiency on the DU25 airfoil surface is wider than that of the NACA63-215 airfoil. Owing to the fact that the impact area of incoming droplets on the DU25 airfoil surface is wider than that on the NACA63-215 airfoil, the distribution range of droplet collection efficiency on the DU25 airfoil surface is wider than that of the NACA63-215 airfoil. In fact, a wide distribution range of droplet collection efficiency on DU-series airfoils was reported by Wang et al. [22], and a narrow distribution range of droplet collection efficiency on NACA63-series airfoils was reported by Guiet et al. [32]. In addition, with increasing angle of attack, there is droplet impingement on the trailing edge of airfoils. This is due to the fact that the discrepancy in airfoils can lead to a change in droplet flow characteristics.

After understanding the droplet collection efficiency on various airfoil surfaces, the next step is to explore the influence of the angle of attack on ice distribution characteristics. Figure 8 and Figure 9 depict the ice thickness distribution on DU25 and NACA63-215 airfoil surfaces at different angles of attack (0–8°) for an ambient temperature of 258 K. For the upper surface, when the angle of attack increases from 0° to 8°, the peak ice thickness on the leading edge of the DU25 and NACA63-215 airfoils decreases from 5.64 to 4.78 mm and from 5.91 to 4.01 mm, respectively. It is clear that the peak ice thickness on the NACA63-215 airfoil (32.1% reduction) decreases more dramatically than that of the DU25 airfoil (15.2% reduction). At an angle of attack of 0°, the peak ice thickness on the NACA63-215 airfoil is larger than that of the DU25 airfoil. However, at an angle of attack of 8°, the peak ice thickness on the NACA63-215 airfoil is smaller than that of the DU25 airfoil. For the lower surface, as the angle of attack increases from 0° to 8°, the peak ice thickness on the leading edge of the DU25 and NACA63-215 airfoils varies in the range of 5.75 to 6.53 mm (13.6% increase) and 5.82 to 6.68 mm (14.8% increase), respectively. The peak ice thickness on the lower surface of the DU25 airfoil is smaller than that of the NACA63-215 airfoil. It is evident that the peak ice thickness on the lower surface of airfoils increases, which is in contrast with that on the upper surface. It is noteworthy that the peak ice thickness on the trailing edge of the NACA63-215 airfoil (0.18 mm) is greater than that of the DU25 airfoil (0.14 mm).

On the other hand, for the leading edge of airfoils, the ice distribution range on the upper surface of the DU25 airfoil is almost the same as that of the NACA63-215 airfoil at an angle of attack of 0°. At an angle of attack of 8°, the ice distribution range on the upper surface of the DU25 airfoil is broader than that of the NACA63-215 airfoil, whereas the ice distribution range on the lower surface of the DU25 airfoil is narrower than that of the NACA63-215 airfoil. This is due to the fact that different airfoils and angles of attack can cause non-uniform droplet collection efficiency, freezing into various ice shapes under icing conditions.

3.3. Coupling Effect of Ambient Temperature and Angle of Attack

After analyzing the influence of a single factor on icing characteristics, the influence of coupled factors on the peak ice thickness on the surfaces of the DU25 and NACA63-215 airfoils was further explored, including ambient temperature and angle of attack.

Figure 10a illustrates the peak ice thickness on the surfaces of the DU25 airfoil under various icing conditions. It was observed that the peak ice thickness changes more significantly at ambient temperatures of 258–268 K than at ambient temperatures of 248–258 K. This is due to the discrepancy in the droplet freezing rate caused by ambient temperature. In addition, the angle of attack has a much greater effect on the peak ice thickness than ambient temperature. This is ascribed to the fact that the influence of the droplet collection rate caused by angles of attack is greater than that of the droplet freezing rate caused by ambient temperature.

Figure 10b illustrates the peak ice thickness on the surfaces of the NACA63-215 airfoil under various icing conditions. It is obvious that the peak ice thickness changes more significantly at ambient temperatures of 258–268 K than at ambient temperatures of 248–258 K at an angle of attack of 0°. At ambient temperatures of 258–268 K, due to the fact that the freezing rate of droplets is not 100%, different ambient temperatures can affect the droplet freezing rate, and unfrozen water film formed on the surface of the airfoil interacts with ice accretion, leading to significant variation in ice thickness. However, at ambient temperatures of 248–258 K, owing to the fact that the freezing rate of droplets is almost 100%, ice thickness is not significantly affected by ambient temperatures. This phenomenon is similar to the reference [20]. However, with an increase in the angle of attack, the peak ice thickness changes more significantly at ambient temperatures of 248–258 K than at ambient temperatures of 258–268 K at an angle of attack of 8°. This is because the effect of the droplet collection rate caused by angles of attack is greater than that of the droplet freezing rate caused by ambient temperature. Moreover, with an increase in the angle of attack, the peak ice thickness changes almost linearly. The angle of attack has a much greater effect on the peak ice thickness than ambient temperature. This is because the angle of attack affects the droplet flow trajectory, and different angles of attack can lead to large discrepancies in ice thickness due to the variation in droplet collection efficiency. On the other hand, ambient temperature affects the freezing rate of droplets, especially under rime ice conditions, due to the almost 100% freezing rate; the thickness of ice accretion is minimally affected by ambient temperature.

4. Conclusions

In this study, a numerical simulation was used to investigate the effects of ambient temperature and angle of attack on the ice distribution characteristics of different airfoil surfaces for a wind turbine. The following conclusions were deduced:

• For the upper surface, when the ambient temperature increases from 248 K to 268 K, the peak ice thickness on the leading edge of the DU25 and NACA63-215 airfoils is diminished by 2.1% and 3.2%, respectively.
• For the lower surface, the peak ice thickness on the leading edge of the DU25 and NACA63-215 airfoils drops by 4.8% and 6.1%, respectively.
• For the leading edge of airfoils, the ice distribution range on the upper surface of the DU25 airfoil (0–3.07 mm) is wider than that of the NACA63-215 airfoil (0–1.91 mm), while the ice distribution range on the lower surface of the DU25 airfoil (0–12.13 mm) is narrower than that of the NACA63-215 airfoil (0–15.18 mm).
• For the trailing edge of airfoils, the ice distribution range of the DU25 airfoil (80.55–99.36 mm) is broader than that of the NACA63-215 airfoil (85.03–99.36 mm).
• For the upper surface, when the angle of attack increases from 0° to 8°, the peak ice thickness on the NACA63-215 airfoil (32.1% reduction) reduces more dramatically than that of the DU25 airfoil (15.2% reduction).
• For the lower surface, the peak ice thickness on the surface of the DU25 airfoil is smaller than that of the NACA63-215 airfoil.
• At an angle of attack of 0°, the ice distribution range on the upper surface of the DU25 airfoil is almost the same as that of the NACA63-215 airfoil.
• At an angle of attack of 8°, the ice distribution range on the upper surface of the DU25 airfoil (0–1.05 mm) is broader than that of the NACA63-215 airfoil (0–0.675 mm), whereas the ice distribution range on the lower surface of the DU25 airfoil (0–17 mm) is narrower than that of the NACA63-215 airfoil (0–20 mm).
• The angle of attack has a much greater effect on the peak ice thickness than ambient temperature.

In our follow-up study, the effect of airfoil design characteristics on the ice accumulation patterns of wind turbine blades will be explored, including camber and thickness. Afterward, wind turbine blade designs will be optimized, which can provide a reference for the anti/de-icing design of wind turbines.