Experimental Study on the Influence of Groove-Flap and Concave Cavity on the Output Characteristics of Vertical Axis Wind Turbine

Abstract

To address the low wind energy utilization efficiency of vertical axis wind turbines (VAWTs) and enhance their engineering applicability, cavity and groove-flap structures were incorporated into turbine blades. Numerical simulations were performed to optimize these configurations, followed by wind tunnel experiments investigating output power variations of three VAWT types under different wind speeds at installation angles of 0°, 2°, 4°, and 6°. The Omega criterion was employed to comparatively analyze vortex evolution patterns at the leading and trailing edges for installation angles of 0°, 3°, and 5°. Experimental results demonstrated nonlinear growth in output power with increasing wind speed and rotational velocity, with groove-flap VAWTs exhibiting superior performance. The optimal installation angle was identified within 2.5–3.5°, where appropriate angles reduced adverse pressure gradients, delayed boundary layer separation, and mitigated vortex shedding effects. Excessive angles induced vortex accumulation and wake disturbances, compromising flow field stability. This study provides critical insights for optimizing VAWT aerodynamic performance through structural modifications and installation angle adjustments.

1. Introduction

With the construction of smart cities, urban wind energy has become an important factor in the development of smart cities in the future, especially in windy areas with low solar radiation intensity. VAWTs have become a research hotspot for urban applications due to their advantages, such as omnidirectional capability, scalability, robustness, and low cost. Therefore, it is particularly critical to study vertical axis wind turbines suitable for urban environments [1,2,3,4].

Researchers have proposed many effective flow control technology methods. These methods can be divided into active control and passive control. Active control is mainly studied by plasma excitation [5,6,7,8], movable groove-flap [9], and movable Gurney flap [10]. Although the active control strategy can improve the aerodynamic performance of the vertical axis wind turbine in numerical simulation, it is not easy to design, operate, and implement in practical applications and is not suitable for small vertical axis wind turbines in cities.

The passive control techniques are mainly used to enhance the fluid attachment ability and delay the flow separation by changing the airfoil structure. There are mainly methods such as setting bionic nodules [11], porous media [12], rough belts, cavities, and groove-flaps. Tariq et al. used a genetic algorithm combined with a Gaussian process regression model to find the best cavity shape at the stall angle of attack and determined the effectiveness of the eddy current capture ability [13]. Vuddagiri et al. found that a circular cavity near the trailing edge provides better aerodynamic efficiency than other positions [14]. Sobhani et al.’s research suggests that the concave cavity structure can enhance the aerodynamic performance of wind turbines [15]. At TSR = 2.6, the wind energy utilization rate is approximately 18%higher than that of the clean wind turbine [15]. Olsman et al. conducted a two-dimensional numerical simulation of airfoils with recessed cavity structures under low Reynolds number conditions [16]. Their research revealed that the recessed cavity structure would reduce the vortices causing downstream flow separation. Compared with airfoils without the recessed cavity structure, those with the cavity structure would achieve a higher lift-to-drag ratio [16]. Fatehi et al. used a genetic algorithm to optimize the concave cavity structure with the goal of achieving the optimal lift-to-drag ratio [17]. The experimental results demonstrated that the lift-to-drag ratio increased by 31%at an angle of attack (AOA) of 14 degrees and by 57%at an AOA of 20 degrees due to the application of the optimized cavity [17]. Liu et al. conducted a two-dimensional numerical simulation on the NACA-4415 airfoil [18]. Their research revealed that the grooves could eliminate the LSBs formed on the suction surface that result in unsteady vortex shedding [18]. Ibrahim et al. set up cavity structures on the suction surface of the airfoil [19]. The research shows that the cavity structure can enhance the flow control ability of the boundary layer on the suction side of the blade under large angles of attack, especially at low TSRs [19]. Bianchini et al. find that the proper GF configuration can provide a notable increase in the aerodynamic performance, especially at medium-low TSRs [20]. Syawitri et al. used a hybrid model based on the transitional shear-stress transport turbulence model to optimize the geometry of the Gane flap on the VAWT [21]. The results confirmed that different optimal Gane flap geometries have different performances within different TSRs, and the Gane flap performs well at low TSRs [21]. Haitian et al. studied the influence of the Gane flap on vertical-axis wind turbines and the effect of solidity on the two structural types (exposed type and crater type) of SB-VAWT [22]. Ismail et al. added a combination of semi-circular dimples and a Gurney flap at the lower surface of the NACA-0015 airfoil [23]. From the study, it can be concluded that the average tangential force can be increased by approximately 35%in the steady-state case and 40%in the oscillating case (at each revolution) by utilizing an optimized combination of a Gurney flap and a semi-circular inward dimple [23].

The above researchers studied the output power of the blade after adding different structures through wind tunnel experiments but did not consider the relationship between output power and installation angle and rotational speed, nor did they analyze the drag reduction mechanism of the wind turbine based on the installation angle. Therefore, this paper adopted a method combining numerical simulation and wind tunnel experiments to conduct the research. Firstly, two wing-type optimization structures were proposed: the cavity structure and the groove-flap structure. The wing type structure optimization was carried out using numerical simulation; secondly, after determining the optimal structures of the two wing types, wind tunnel experiments were used to compare the output power changes in the optimal cavity wind turbine and the optimal groove-flap wind turbine under different wind speeds at different rotational speeds and installation angles compared to the clean wind turbine; finally, the optimal installation angles of the clean wind turbine, the cavity wind turbine, and the groove-flap wind turbine were found using numerical simulation, and the influence of the installation angle on the drag reduction mechanism of the wind turbine was analyzed from the perspective of vorticity, providing theoretical and experimental references for the application of small vertical-axis wind turbines.

2. Wind Turbine Model

2.1. Geometric Model

This study employs a NACA0012 symmetrical airfoil as the research subject, incorporating a blunt trailing edge modification based on the clean vertical axis wind turbine configuration. The baseline parameters of the reference wind turbine are detailed in

Table 1. Basic parameters of VAWT.

Parameter	Numerical Value
Number of blades	3
Solidity	0.73
Revolution speed/rpm	180–300
Airfoil chord length c/mm	210
Wind wheel diameter D/mm	860
Wind wheel stretch H/mm	1000

. To enhance aerodynamic characteristics, specially designed cavity structures and groove-flap configurations have been integrated into the upper and lower airfoil surfaces, respectively. The cavity structure is defined by non-uniform rational B-spline (NURBS) curves. The groove-flap structure is determined by selecting appropriate parameters such as the height of the Gurney flap, the distance between the Gurney flap and the tail edge, and the diameter of the groove.

The main performance parameters of vertical axis wind turbines include tip speed ratio ($\lambda$), air density ($\rho$), incoming wind speed ($U_{\infty }$), and rotor power coefficient ($C_P$):

$$\lambda ={\displaystyle \frac{\omega R}{U_{\infty }}}$$

(1)

$$C_P={\displaystyle \frac{M\omega }{0.5\rho {U_{\infty }}^{2}A}}$$

(2)

2.2. Computational Domain and Grid Division

To ensure computational model reliability, numerical simulations were conducted using ANSYS Fluent 2022 R1 software. The two-dimensional computational domain and fluid boundary conditions are illustrated in Figure 1. The fluid domain comprises blade mesh refinement zones (G1, G2, G3), a rotating zone (G4), and the main fluid domain (G5). The mesh refinement zones enhance resolution of blade surface flow dynamics, while the rotating zone governs blade rotational speed in Fluent. Interfaces were established between refinement/rotating zones and rotating/fluid domains to enable inter-region data transfer. All three blades were configured as no-slip walls. The inflow wind speed was set to 10 m/s during simulations.

Mesh generation is a critical component of computational fluid dynamics (CFD) methodology. In this study, ICEM software 2022 is employed to generate the mesh. The blade mesh is refined with a Y+ value set to 1, and the first-layer grid height near the wall is maintained at 0.0029 mm. The grid elements progressively increase with a growth factor of 1.05, as illustrated in Figure 2.

During the operation of the vertical axis wind turbine, the Mach number remains below 0.3, allowing the fluid to be treated as incompressible. The Transition SST (TSST) turbulence model is selected for turbulence closure. The pressure-based transient solver is employed, utilizing the SIMPLEC algorithm for pressure-velocity coupling and a second-order upwind scheme for spatial discretization. The time step for the unsteady simulation is dynamically adjusted based on the tip speed ratio (TSR) to ensure that each time step corresponds to 1° of azimuthal rotation, with 50 iterations performed per time step.

2.3. Reliability Demonstration

In the numerical simulations, three mesh schemes with grid counts of 636,000, 850,000, and 1,200,000 were designed. The C_P values were numerically observed under different tip speed ratios. The C_P values of the turbine blade stabilized during the seventh revolution, and data from the eighth revolution were extracted for grid verification. As shown in Figure 3, the C_P values of the 850,000-cell mesh exhibited smaller deviations compared to those of the 1,200,000-cell mesh and aligned more closely with reference values from literature. Consequently, the 850,000-cell mesh was adopted as the baseline grid for subsequent numerical calculations.

2.4. The Determination of the Cavity Structure

The cavity initiation point is positioned at 75% chord length. The start(S) and terminal(T) points of the cavity are fixed, separated by 10%c in the axial direction. Three movable control points (M1, M2, and M3) are designated along the NURBS curve to parametrize the cavity geometry. Structural optimization is achieved by relocating these control points. A Latin Hypercube Sampling (LHS) technique is employed to systematically generate 80 parameter combinations through uniform sampling within the design space. As illustrated in Figure 4, the blue, magenta, and purple point clusters correspond to the feasible domains of M1, M2, and M3, respectively. Numerical annotations following S and T denote their chordwise coordinates, while those adjacent to M1–M3 specify permissible coordinate ranges. Black directional arrows indicate permissible adjustment vectors for the control points.

Eighty datasets within the parameter range were defined with the lift-to-drag ratio at the stall angle of attack as the objective function. Figure 5 presents a comparison of the lift coefficient (C_L) and drag coefficient (C_D) for the clean blade under varying angles of attack. The left panel illustrates variations from 2° to 13°, while the right panel details changes starting at 12° with 0.1° increments.

From Figure 5, it is observed that the blade begins to stall at an angle of attack of 13°. Consequently, numerical simulations were conducted for the aforementioned 80 datasets at this stall angle. After completing the simulations, the top three concave cavity structures with the highest lift-to-drag ratios were selected and labeled A, B, and C. The schematic diagrams of cavities A, B, and C are shown in Figure 6; the cavity structure A protrudes towards the leading edge of the airfoil, the cavity structure B takes on an elliptical shape, and the cavity structure C protrudes towards the trailing edge of the airfoil.

Unsteady numerical simulations were then performed on these optimized configurations, as illustrated in Figure 7. As can be seen in Figure 7, the cavity configuration enhances the airfoil’s power coefficient (wind energy utilization rate), albeit with limited magnitude. Among the three cavity geometries, configuration A demonstrates superior aerodynamic performance. At tip speed ratios (TSRs) of 1.2, 1.4, 1.6, 1.8, and 2.0, configuration A achieves relative power coefficient improvements of 3.88%, 10.16%, 3.88%, 1.28%, and 2.26%, respectively, compared to the baseline wind turbine.

2.5. Determination of Groove-Flap Airfoil

The key geometric parameters of the groove-flap configuration include Gurney flap height (h), trailing-edge offset distance (L), groove diameter (d), and Gurney flap width (T), as illustrated in Figure 8. Given the numerous influencing factors, conducting simulations for all experimental combinations would be computationally intensive and impractical for isolating individual effects on VAWT performance. To address this, an orthogonal experimental design was adopted, replacing full factorial tests with representative sampling. Range analysis was then applied to evaluate the impact of each factor on the reference metrics, enabling the identification of an optimal parameter combination. The orthogonal test combination is shown in

Table 2. Orthogonal test combination.

Numbering	Factor
	A	B	C
Case-1	1.00%c	5.00%c	1.00%c
Case-2	1.00%c	10.00%c	1.50%c
Case-3	1.00%c	15.00%c	1.25%c
Case-4	1.25%c	5.00%c	1.50%c
Case-5	1.25%c	10.00%c	1.25%c
Case-6	1.25%c	15.00%c	1.00%c
Case-7	1.50%c	5.00%c	1.25%c
Case-8	1.50%c	10.00%c	1.00%c
Case-9	1.50%c	15.00%c	1.50%c

.

The Gurney flap height (h), distance from the trailing edge (L), and groove diameter (d) were selected as parameters A, B, and C in the orthogonal experimental design. Considering the actual airfoil chord length (c) and referencing existing literature, the parameters were defined as follows:

(1) h: Gurney flap heights of 1.00%c, 1.25%c, and 1.50%c;

(2) L: Distances from the trailing edge set to 5.00%c, 10.00%c, and 15.00%c;

(3) d: Groove diameters equal to the Gurney flap height (h), thus also adopting 1.00%c, 1.25%c, and 1.50%c;

(4) T: Gurney flap width fixed at 1 mm.

Figure 9 compares the power coefficient (wind energy utilization rate) of the clean airfoil at tip speed ratios (TSRs) of 1.2, 1.4, 1.6, 1.8, and 2.0 with nine test cases derived from orthogonal experimental design. The integration of groove flaps significantly enhances aerodynamic performance, exhibiting a nonlinear increase in power coefficient with rising TSR until reaching a critical ratio, beyond which gradual performance degradation occurs. Test case 7 demonstrates superior overall improvement, achieving a peak power coefficient of 0.14 at TSR 1.6—a 14.3% enhancement relative to the clean airfoil. Consequently, the groove-flap configuration from test case 7 was selected for implementation in the experimental wind turbine blade.

To further analyze the influence of the three aforementioned factors on the power coefficient. The arithmetic mean and range values of each factor were calculated to identify the most influential parameter. According to range analysis principles, a larger range value indicates a stronger effect of the corresponding factor on the objective function. As summarized in

Table 3. Arithmetic mean and range of each factor at different tip-speed ratios.

Numbering	1.2			1.4			1.6			1.8			2.0
	A	B	C	A	B	C	A	B	C	A	B	C	A	B	C
K1	0.235	0.246	0.239	0.313	0.324	0.304	0.371	0.396	0.368	0.398	0.411	0.384	0.378	0.381	0.366
K2	0.230	0.242	0.235	0.297	0.313	0.309	0.353	0.396	0.371	0.386	0.382	0.395	0.364	0.367	0.373
K3	0.248	0.226	0.242	0.311	0.283	0.308	0.386	0.345	0.371	0.376	0.367	0.380	0.363	0.357	0.366
$\overline{K1}$	0.078	0.082	0.080	0.104	0.108	0.101	0.124	0.132	0.124	0.133	0.137	0.128	0.126	0.127	0.122
$\overline{K2}$	0.077	0.081	0.078	0.099	0.104	0.103	0.118	0.123	0.124	0.129	0.127	0.132	0.121	0.122	0.124
$\overline{K3}$	0.083	0.075	0.080	0.104	0.094	0.103	0.129	0.115	0.124	0.125	0.122	0.127	0.121	0.119	0.122
R	0.006	0.007	0.002	0.005	0.014	0.001	0.011	0.017	0.001	0.007	0.015	0.005	0.005	0.008	0.002

, K1, K2, and K3 represent the sum of metric values for Factors A, B, and C, respectively; $\overline{K1}$, $\overline{K2}$ and $\overline{K3}$ denote their arithmetic means, and R corresponds to the range (difference between maximum arithmetic and minimum arithmetic mean values) for each factor. The wind energy utilization values for the nine cases determined through orthogonal experiments are presented in the Appendix A.

Analysis of Table 3 reveals that among factors A and C, the range value R of factor B is the largest. From the low tip speed ratio of 1.2 to the high tip speed ratio of 2.0, the range values of factor B are 0.010, 0.014, 0.017, 0.015, and 0.008, respectively. Therefore, factor B, namely the angle of the Gurney flap relative to the tail edge position, has the greatest impact on improving the utilization rate of wind energy.

3. Experimentation

3.1. Laboratory Equipment

The experiment utilized an open-circuit B1/K2 low-speed wind tunnel with a total length of 24.6 m and a test section length of 20.8 m. The tunnel’s centerline height was 1.7 m. A digital frequency conversion system (0–50 Hz) regulated inflow velocity, achieving stable wind speeds ranging from 0 to 20 m/s. A Fluke Norma 5000 high-precision power analyzer (Fluke Corporation, Everett, WA, USA) and a programmable DC load bank were employed to monitor turbine output power and adjust rotational speed, while a hot-wire anemometer calibrated wind speed.

The wind turbine structure employs bolted connections, where blades are fastened to support rods, support rods to the main shaft, the main shaft to the base frame, and the base frame to the ground foundation using bolts. All components—including bolts, washers, support rods, and others—were identical in mass throughout the experimental setup. The generator used had a rated power of 200 W. The turbine blades, designed through numerical simulations and fabricated via 3D printing technology, measure 1000 mm in height and are composed of white resin. To prevent stress concentration and structural failure during rotation, critical regions near bolt holes were reinforced with solid material, while the remaining blade sections were hollowed to reduce overall mass. The primary experimental equipment and fully assembled turbine configuration are illustrated in Figure 10.

3.2. Experimental Scheme

A total of three types of wind turbines were used in the experiment, namely the clean wind turbine, the cavity wind turbine, and the groove-flap wind turbine. The blades used for the cavity wind turbine and the groove-flap wind turbine were derived from the best cavity structure A in Section 2.4 and the optimal groove-flap structure Case-7 in Section 2.5. Four data points were measured for each group of blades at four installation angles: 0°, 2°, 4°, and 6°. The installation angles were changed by adjusting the support rods and the blades.

To ensure operational safety, all three wind turbines maintained a rotational speed of 300 rpm. Data acquisition occurred at 30-revolution intervals, with each recording session lasting 2 min followed by a 1-min stabilization period. Prior to each measurement cycle, anemometric calibration was performed to mitigate ambient temperature effects on wind tunnel velocity.

3.3. Experimental Results and Analysis

The output power of the clean wind turbine, groove-flap augmented turbine, and cavity wind turbine under varying wind speeds demonstrates analogous behavioral patterns. All three VAWT configurations exhibit similar nonlinear power responses to blade initial pitch angle variations: the output power initially increases with pitch angle adjustments, reaches an optimal value, and subsequently undergoes abrupt reduction.

Under low wind speed conditions (6 m/s), as shown in Figure 11, all three turbine configurations exhibited comparatively low power outputs. However, performance benchmarking revealed that the cavity wind turbine demonstrated superior operational efficacy under identical conditions. With a blade pitch angle of 2°, the cavity configuration achieved power output enhancements of 0.98%, 6.28%, 14.52%, 24.24%, and 15.49%relative to the clean wind turbine at rotational speeds of 180 rpm, 210 rpm, 240 rpm, 270 rpm, and 300 rpm, respectively. Notably, progressive pitch angle increases correlated with a monotonic decline in power generation efficiency for the cavity design.

At different rotational speeds, the output power at an installation angle of 6° is much lower than that at an installation angle of 0°. Taking the clean wind turbine as an example, at 8 m/s, as shown in Figure 12, the clean wind turbine decreased by 18.50%, 89.72%, 59.12%, 59.23%, and 63.03%, respectively, at different rotational speeds (180 rpm, 210 rpm, 240 rpm, 270 rpm, and 300 rpm).

At 10 m/s as shown in Figure 13, the groove-flap wind turbine exhibited optimal power performance. With a 2° pitch angle configuration, this modified turbine outperformed the clean one by 5.86%, 10.30%, 16.06%, 24.78%, and 31.87%across the same rotational speed range.During the experiment, it was also observed that with the increase of the installation angle, the blade start-up showed a pendulum-shaking phenomenon. This dynamic instability not only reduces the aerodynamic efficiency of the wind turbine but also may pose a potential threat to the structural strength and service life of the blade.

Simultaneously, under low rotational speed conditions (e.g., 180 rpm), both the groove-flap wind turbine and cavity wind turbine exhibit negligible aerodynamic performance improvements, maintaining an output power of approximately 5 W. As rotational speed increases, the design advantages of these optimized blade configurations gradually manifest. This demonstrates that blade structure optimization effectively enhances the overall efficiency of wind turbines across varying operational conditions.

4. Numerical Simulation Analysis

4.1. Analysis of Numerical Simulation Results and Experimental Results

To further investigate the optimal pitch angle and analyze the aerodynamic characteristics of three VAWTs under varying pitch configurations, this study employs numerical simulations to examine power coefficient variations with pitch angle at a wind speed of 10 m/s.

As depicted in Figure 14, peak power coefficients for all three VAWT configurations occur within a pitch angle range of 2.5–3.5°.

Due to the use of small-scale generators, the output power of all three vertical-axis wind turbines is relatively low values low in magnitude, with a maximum of approximately 30 W, and even lower values under low wind speeds. However, comparative analysis between experimental results and numerical simulations reveals that the optimized VAWT exhibits a significantly higher power output increase than predicted by the simulations. This discrepancy can be attributed to three primary factors.

First, in the numerical simulation, all boundary conditions are idealized. For example, the incoming wind speed remains consistent with the initial settings when interacting with the blades. In contrast, during experiments, the wind turbine is positioned 0.5 m downstream from the wind tunnel outlet. The diffusion effect of the airflow at the tunnel exit induces local acceleration phenomena above and below the tunnel, which amplifies localized aerodynamic loads on the blades, thereby marginally enhancing the output power.

Second, during experimental testing, the generator operates under load-controlled conditions where blade rotational speed is regulated by adjusting the electrical load. This setup simplifies the fluid-blade interaction to a unidirectional airflow contact process, deviating from the turbine’s actual operational dynamics.

Third, numerical simulations typically assess turbine stability by monitoring periodic fluctuations of a selected variable. A system is deemed stable when the variable oscillates periodically with a relative error below 3%over consecutive cycles. Post-stabilization, simulations follow predefined mathematical models with reduced data variability, yielding higher theoretical consistency. This inherent stability, however, limits the ability to capture dynamic fluctuations. In contrast, experimental data inherently incorporate external disturbances under non-idealized conditions, providing a more direct representation of real-world turbine behavior.

Therefore, the integrated approach combining experimental results with numerical simulations provides critical insights for further optimization of wind turbine design. Experimental validation bridges the gap to real-world operational scenarios, while numerical simulations elucidate the inherent principles of aerodynamic design under idealized conditions and enable reliable prediction of performance variations. The synergy of these two methodologies facilitates a comprehensive evaluation of wind turbine performance and advances aerodynamic optimization efforts.

4.2. Analysis of Wind Turbine Vorticity

To elucidate the impact of pitch angle on aerodynamic performance, flow field characteristics under different pitch angles were analyzed through vortex dynamics using the Omega vortex identification criterion. This criterion demonstrates superior capability in capturing weak vortices, fragmentary micro-vortices, small-scale vortex shedding, and secondary flow structures [24].

Since the working area of the vertical axis wind turbine is primarily concentrated in the upwind region, the vorticity variation cloud diagrams were specifically observed under the following conditions at 240 rpm: the groove-flap wind turbine blade 1 at 30° azimuth angle, the clean wind turbine blade 1 at 60° azimuth angle, and the cavity wind turbine blade 1 at 120° azimuth angle.

As shown in Figure 15, the Omega vorticity cloud map of the groove-flap wind turbine blade 1 at an azimuth angle of 30° indicates that the vortex intensity in the flow field gradually increases from blue to red. Figure 15a shows the overall changes in the vorticity cloud map of the groove-flap wind turbine from left to right at installation angles of 0°, 3°, and 5°. Figure 15b shows the changes in the vorticity cloud map of blade 2 of the trough-flap wind turbine at azimuth angle 150° from left to right.

At an installation angle of 0°, a moderate-intensity vortex core is generated at the trailing edge of Blade 1. When the installation angles are 3° and 5°, the vortex cores at the trailing edge of Blade 1 are smaller and exhibit lower intensity. Blade 2 of the vertical axis wind turbine with these three installation angles produces vortices of varying quantities and intensities at both the leading and trailing edges, indicating that the flow near the blade surface is significantly influenced by shear stress and pressure gradients. At a 0° installation angle, the leading edge separation vortex shows a large size and high intensity, suggesting that a pronounced adverse pressure gradient develops on the windward side. This gradient leads to more severe flow separation near the leading edge, thereby intensifying pressure resistance on the blade surface. In contrast, at installation angles of 3° and 5°, the leading edge separation vortices are smaller in size and weaker in intensity, implying a mitigation of the adverse pressure gradient under these configurations. However, as the installation angle increases, the shedding vortices in the downwind region become concentrated near the trailing edge of Blade 1 and the leading edge of Blade 3. Notably, the accumulation of shedding vortices is most pronounced at a 5° installation angle, which amplifies flow field instability and results in energy loss within the wake region.

Figure 16 displays the Omega vorticity cloud map of the clean wind turbine Blade 1 at an azimuth angle of 60°. Overall, the shedding vortex accumulation in the downwind region at an installation angle of 3° is less pronounced, ensuring the stability of the flow field structure. Blade 2 is positioned at an azimuth angle of 180°. When the inflow passes from the trailing edge to the leading edge along the blade surface, all three installation angles generate substantial leading-edge separation vortices at the leading edge. This phenomenon causes uneven pressure distribution on the blade surface, exacerbates aerodynamic drag, and reduces blade lift. Specifically, the installation angle of 0° exhibits the largest vortex core size and the highest intensity, corresponding to the poorest aerodynamic efficiency. At the trailing edge of the blade, the trailing-edge vortex structures demonstrate relatively low intensity. The presence of these trailing-edge vortices indicates intense energy exchange during flow passage through the trailing edge. At an installation angle of 5°, the increased installation angle leads to further expansion of vorticity distribution in the wake region, thereby inducing wake instability.

Figure 17 shows the Omega vorticity cloud map of the cavity vertical axis wind turbine Blade 1 at an azimuth angle of 120°. Similarly, at an installation angle of 0°, the leading edge of Blade 1 exhibits large-scale, high-intensity separation vortices and small fragmentary vortices. This phenomenon is primarily attributed to the small angle between the incoming flow velocity direction and the blade surface, resulting in a sharp variation in the leading-edge pressure gradient and triggering premature flow separation. The trailing-edge vortex structures at the blade trailing edge follow an evolution pattern similar to that observed in the clean wind turbine.

Furthermore, as can be seen from Figure 15a, Figure 16a, and Figure 17a, the clean wind turbine, the cavity wind turbine, and the groove-flap wind turbine all mainly have their shedding vortices concentrated in the downwind area. This phenomenon primarily occurs because vortex shedding takes place when the blade transitions from the upwind to the downwind region. The vortices generated by shedding interact with those originating from the downwind region, resulting in vortex counteraction. The interaction between these vortices induces pressure pulsations on the blade surface, which further explains why the primary working area of the wind turbine is concentrated in the upwind region.

5. Conclusions

In this study, cavity and groove-flap structures were respectively added into the blades of an H-type vertical axis wind turbine. The optimal structural configuration and geometry were determined using NURBS curves and orthogonal experimental design. Experimental investigations were conducted to evaluate the effects of different blade configurations on output power and initial installation angles. Comparative analysis of vortex evolution patterns at the leading and trailing edges under installation angles of 0°, 3°, and 5° was performed based on the Omega criterion, providing insights for future urban applications of VAWTs. The key findings are summarized as follows:

(1) Optimal Installation Angle: The three VAWT configurations exhibit an optimal installation angle range of 2.5–3.5°. Both cavity and groove-flap structures demonstrate enhanced aerodynamic performance with increasing wind speed and rotational speed.

(2) Pressure Gradient Mitigation: Appropriate installation angles reduce the adverse pressure gradient on the blade surface, delay boundary layer separation, and diminish the impact of shedding vortices on the flow field—particularly in reducing the intensity and scale of leading-edge shedding vortices.

(3) Excessive Angle Effects: Overly large installation angles induce shedding vortex accumulation in the downwind region. Concurrently, vorticity distribution at the blade trailing edge expands, triggering wake disturbances and compromising flow field stability.