Experimental and numerical investigation of Zephyr-type wind turbine

Abstract

The search for more environmentally friendly energy sources has been prompted by growing environmental concerns. In this regard, wind energy can be an alternative viable source of energy for world electricity demand supply. Despite of their benefits in terms of simplicity of manufacture, independency on wind direction and good starting ability in turbulent flow, Savonius wind turbines as a vertical axis wind energy converter are not recommended for large-scale power generation because of their poor performances. This research emphasizes on the performance improvement of a Zephyr wind rotor (ZWR). Experimental tests were conducted in a wind tunnel for different Reynolds numbers. Maximum power coefficient of 0.074 was recorded for Re = 158,000 corresponding to V_∞ = 10 m s⁻¹ at a tip speed ratio of λ = 0.99. Numerical study was carried out with the use of Ansys Fluent 17.0 software through transient 3D simulations investigating the optimization of the ZWR rotor blades and stator vanes with the aim of performance betterment. Maximum power coefficient of 0.168 was found with 8-bladed ZWR with 12 stator vanes. Aerodynamic properties were also improved.

Keywords

Zephyr wind rotor wind tunnel rotor blades stator vanes performance betterment aerodynamic properties

Introduction

As a result of the negative environmental impact and exhaustible supply of fossil fuels, global efforts have shifted to renewable and environmental friendly source of energy, primarily wind power, to meet the increasing electricity demand. Wind turbines defined as wind converters are categorized into horizontal (HAWT) and vertical axis (VAWT) ones. Depending on the driving force, VAWT are classified into two types: the drag force-based Savonius rotor and the lift force-based Darrieus rotor (Shende et al., 2022). The Savonius rotors offer a number of advantages over the HAWT and the Darrieus rotors, even though their power efficiency is lower. They show a lesser manufacturing and maintenance cost, a far simple design and especially a good starting capability to operate under turbulent flow independently to the wind direction. Its conventional design consists of two semi-cylindrical blades positioned from either side of a vertical shaft, giving the impression of an “S” form. The idea behind this is to introduce wind into the turbine blades.

The aspect and the overlap ratio, the number of blades and stages, the use of end plates as well as the blade shape represent the main geometrical factors affecting the performance of the conventional Savonius wind rotor. It can equally be affected by adding aerodynamic appendages. Numerous studies have shown that parametric optimization has enhanced the Savonius rotor performance both numerically and experimentally. Regarding the aspect ratio, Payambarpour et al. (2020) deduced that when it rises, the torque coefficient increases while the rotating speed decreases because the turbine structure gets heavier. They found that 1.0 was the optimum value. Hassanzadeh and Mohammad Nejad (2019) tested different values of inward and outward overlap ratios. The 0.2 overlapped Savonius rotor highlighted the highest performances. In the same vein, numerically, Patel et al. (2023) examined negative overlap ratios in the interval of [0–0.4]. They found that, for a tip speed ratio of 0.6, the zero overlap ratio (non overlapped rotor) produced the highest power coefficient, which was equivalent to 0.09. Furthermore, the use of end plates was thought to improve the Savonius rotor performance. In a similar spirit, Jeon et al. (2015) experimented with various lower and higher end plate arrangements for Savonius rotors. It was reported that the power coefficient has improved by 36%. Mahmoud et al. (2012) investigated the blades number. They found that because of the significant vortex phenomena that forms around the rotor blades, the rotor performance decreases as the number of blades increases. In comparison to rotors with three and four blades, the two-blade Savonius rotor displayed the highest value of power and torque coefficients. For the stages number, Shamsuddin et al. (2023) claimed that the performance of the double-stage rotor was higher than that of the single stage. The maximum power coefficient measured was 13% higher at 5 m s⁻¹. Four blade shape configurations: the semi-circular, Benech, modified-Bach, and elliptical were identified by Alom and Saha (2019) and subsequently subjected to computational and experimental testing. Numerically, the maximum power coefficient was equal to 0.27, 0.29, 0.30, and 0.34, in that order. As a result, the elliptical profile performed the best. Besides, the blade curve was addressed by Jia et al. (2022). They presented a newly designed two-blade Savonius rotor that makes use of the cubic Bezier curve technique. The enhanced blade was examined and contrasted with the traditional one. It emphasized two advantages. On the one hand, it increased the aerodynamic performance and by 6.87% in the power coefficient. For the use of aerodynamic appendages, numerous designs were looked over in this regard. Deflecting plates were placed ahead the rotor. The investigated designs were set of only one plate (Putri et al., 2019) or even two plates (Mohamed et al., 2021; Saikot et al., 2023). Deflectors (Idrissi et al., 2023; Layeghmand et al., 2020) were also proposed and studied. It was discovered that using those accessories significantly enhanced the Savonius rotor aerodynamic properties and output performances, particularly the torque and power coefficients.

In this regard, a review of researches addressing the Zephyr wind rotor (ZWR) which represents a developed design of Savonius rotors shows a scarcity of studies concerning their performance amelioration. Most of researches focused on the performance improvement of the conventional design of the Savonius wind rotor. Thus, the basic objective of this work is to assess and boost the performance of a ZWR prototype both experimentally and numerically. To do so, experimental tests were conducted in a wind tunnel for different Reynlods numbers. Numerical study was realized with the use of Ansys Fluent software. The optimization involves adjusting the number of stator vanes and rotor blades.

Experimental methodology

Wind tunnel

The wind tunnel presented in Figure 1 was used to determine the aerodynamic features of a ZWR prototype. It is constructed from galvanized iron sheet. It is made up of five main components: a settling chamber in order to shape the flow to be uniform, a contraction cone to speed up the flow, a test section where the prototype was placed, a diffuser to decompress the flow and slower it before it returns to the fan located in a drive section that contains an aspirator used to suck the air to the test section. By altering the frequency of the fan, the air velocity can be managed subsequently the turbulence regime. Thus, under these study, three regimes were tested characterized by an air velocity equal to V = 7 m s⁻¹, V = 9 m s⁻¹ and V = 10 m s⁻¹ corresponding to the Reynolds numbers of Re = 111,000, Re = 143,000 and Re = 158,000, respectively. The test section is made of Plexiglas with size of 790 mm × 400 mm × 400 mm.

Figure 1.

Wind tunnel.

Rotor manufacturing

The ZWR prototype, presented in Figure 2, is a modified design of the conventional Savonius wind rotor. In fact, it is composed of two parts: the stator and the rotor. The stator is composed of 12 removable vanes with two rectangular endplates. Concerning the rotor is made up of four semi cylindrical blades with two circular endplates. Both the rotor blades and the stator vanes were manufactured from galvanized iron with 2 mm of thickness. All geometrical parameters are summarized in Table 1.

Figure 2.

ZWR prototype.

Table 1.

ZWR geometrical parameters.

Parameters	Dimension
Rotor diameter, D	230 mm
Rotor height, h	257 mm
End plate diameter, D_e	240 mm
End plate thickness	7 mm
Shaft length	257 mm
Shaft diameter, s	10 mm
Number of rotor blades	4
Blade radius, r	35 mm
Stator end plate width, w	300 mm
Stator height, H	325 mm
Vane length, l	290 mm
Vane width, W_v	30 mm
Number of stator vanes	12

Experimental set up and equipments

The ZWR prototype was placed at the middle of the test section in a wind flow for the suggested Reynolds numbers subsequently for different wind velocity measured via a hot wire anemometer AM-4204. Initially, the rotor was rotating freely. For different Reynolds number, Static and dynamic tests were carried out on the examined rotor for two configurations: one without the stator vanes and the other with all stator vanes.

Concerning static measurements, they were realized in every incidence angle Θ ranging from 0° to 360° with steps of 30°. In fact, the rotor prototype was stopped in each Θ determined using a gradual disk and the static torque was displayed via a static torque meter. Next, by figuring out the static torque coefficient, the rotor capacity for self-starting was evaluated. All required instruments for static measurements are shown in Figure 3.

Figure 3.

Experimental apparatus for static measurements.

Concerning dynamic measurements, loads were applied to the ZWR to slow down its rotation in a gradual bit by bit. For each load, a digital tachometer was used to estimate the ZWR rotational speed Ω and a dynamic torque meter was deployed to estimate the dynamic torque T with the use of Emperor Software. Indeed, the instantaneous torque was displayed on the used computer interface then its average value was calculated. The software offers a level of accuracy that cannot be reached by manual testing which minimize the experimental errors. Figure 4 depicts the necessary experimental equipments for dynamic measurements.

Figure 4.

Experimental apparatus for dynamic measurements.

Numerical methodology

Several applications have used computational fluid dynamics (CFD) to solve flow difficulties (Abid et al., 2023; Mosbahi et al., 2021a). By solving the governing equations that describe the flow processes, one may accurately anticipate, evaluate, and interpret the wind flow around turbines in aerodynamics. To estimate the flow characteristics around complex structures, ANSYS Fluent software have been used. In actuality, it solves the Navier Stokes equations using the Finite Volume Method (FVM). This approach is particularly useful for simulating the interaction between the rotor blades and the incoming wind flow in aerodynamics.

Computational domain and boundary conditions

The ZWR numerical model adopt for the numerical investigation was designed via Solidworks as shown in Figure 5. It has the same geometrical parameters of that used in experiments but without the stator vanes. It is made of four blades with a blade chord length of d = 35 mm, an overall diameter of D = 230 mm, a height of H = 257 mm two end plates with a diameter equal to D_e = 240 mm.

Figure 5.

ZWR numerical model.

The computational domain, presented in Figure 6, is divided into two parts: a steady domain that represents the test section and a rotating domain, containing the ZWR, represented by a cylinder. These two domains are separated by a sliding surface. For the stationary domain is 400 mm × 400 mm × 790 mm. A velocity inlet boundary condition was defined and set equal to constant velocity of V_∞ = 10 m s⁻¹. An outlet pressure was designated as the atmospheric pressure. With a rotational speed that matched the ZWR rotational speed, a no-slip moving walls condition was applied to the rotor blades.

Figure 6.

Computational domain and BC.

Meshing

The created computational domain was brought in ANSYS meshing interface. For the entire domain, an unstructured mesh with tetrahedral elements was employed. In addition, at the level of the rotating domain, a mesh refinement was accomplished. At the level of the ZWR blades, inflation with twenty prismatic layers and a 1.2 growth rate was produced in order to capture the quick variations of the aerodynamic parameters around the rotor. The non-dimensional parameter y⁺ was set less than 1. The generated grid is presented in Figure 7.

Figure 7.

Generated mesh.

Turbulence modeling and solver setting

Under this study, three-dimensional transient simulations were carried out through ANSYS FLUENT 17.0 to modelize the air flow around the investigated ZWR through resolving the Navier-Stokes equations (Mosbahi et al., 2021b) for Re = 158,000 corresponding to an inlet velocity of V_∞ = 10 m s⁻¹. Semi Implicit Linked Equations known as the (SIMPLE) technique was used with a second order upwind scheme.

The momentum equation (1), along with the continuity equation (2), specify the Navier-Stokes governing equations for a Newtonian fluid.

\frac{\partial ρ}{\partial t} + \frac{\partial (ρ u_{i})}{\partial x_{i}} = 0

(1)

where ρ represents the air density (kg m⁻³), t discloses the time (seconds) and u_i points out the velocity component defined in the x_i coordinate direction x_i = (x,y,z).

\frac{\partial (ρ u_{i})}{\partial t} + \frac{\partial (ρ u_{i} u_{j})}{\partial x_{j}} = - \frac{\partial p}{\partial x_{i}} + \frac{\partial}{\partial x_{j}} [μ (\frac{\partial u_{j}}{\partial x_{j}} + \frac{\partial u_{j}}{\partial x_{i}} - \frac{2}{3} δ_{ij} \frac{\partial u_{i}}{\partial x_{i}})] + \frac{\partial (- ρ \bar{u'_{i} u'_{j}})}{\partial x_{j}} + F_{i}

(2)

where p and F_i expresses, respectively, the pressure (Pa) and the external forces applied (N).

The mathematical expression for the Reynolds stress tensor constituent parts is provided by equation (3).

- ρ \bar{u'_{i} u'_{j}} = μ_{t} (\frac{\partial u_{i}}{\partial x_{j}} + \frac{\partial u_{j}}{\partial x_{i}}) - \frac{2}{3} ρ k δ_{ij}

(3)

where μ_t denotes the turbulent viscosity (Pa.s), x_i, x_j point the Cartesian coordinate, k expresses the turbulent kinetic energy (Pa.s) and $δ_{ij}$ specifies the Chronecker indices.

The turbulence model undertaken in this study was the Shear Stress Transport (SST) k−ω. It has been shown by numerous researchers to be highly capable at modeling wind turbines and accurately predicting flow patterns (Jaohindy et al., 2013; Lajnef et al., 2020; Sharma and Sharma, 2016). In fact, it incorporates the free flow of the k−ω model in the far wall sections with the precision and robustness of the k−ω model in the near wall regions.

The following equation is the expression for the kinetic energy (k) of the turbulence:

\frac{\begin{matrix} \partial (ρ k) \end{matrix}}{\partial t} + \frac{\begin{matrix} \partial (ρ u_{i} k) \end{matrix}}{\partial x_{i}} = \frac{\begin{matrix} \partial \end{matrix}}{\partial x_{i}} [(μ + \frac{μ_{t}}{σ_{k}}) grad (k)] + 2 μ_{t} \frac{\begin{matrix} \partial u_{i} \end{matrix}}{\partial x_{j}} \cdot \frac{\begin{matrix} \partial u_{i} \end{matrix}}{\partial x_{j}} - \frac{2}{3} ρ k \frac{\partial u_{i}}{\partial x_{j}} {δ_{ij}}_{k} - β^{*} ρ k ω

(4)

The specific dissipation rate (ω) is written as follows:

\begin{matrix} \frac{\partial ρ ω}{\partial t} + \frac{\partial (ρ u_{i} ω)}{\partial x_{i}} = \\ \frac{\partial}{\partial x_{i}} [(μ + \frac{μ_{t}}{σ_{ω, 1}}) grad (ω)] + γ_{2} (2 ρ \frac{\partial u_{i}}{\partial x_{j}} \cdot \frac{\partial u_{i}}{\partial x_{j}} - \frac{2}{3} ρ ω \frac{\partial u_{i}}{\partial x_{j}} δ_{ij}) - β_{2} ρ ω^{2} + 2 \frac{ρ}{σ_{ω, 2} ω} \frac{\partial k}{\partial x_{k}} \frac{\partial ω}{\partial x_{k}} \end{matrix}

(5)

where μ presents the dynamic viscosity (Pa.s) and β^*, β₂, $σ_{k}$ , $σ_{ω, 1}$ , $σ_{ω, 2}$ and γ₂ are the constants of the SST k−ω turbulence model.

Naturally, turbulent flow surrounds the ZWR. Under this condition, transient simulations were run. Based on equation (6), the time step (Δt) was appropriately selected in relation to the ZWR rotation properties.

Δ t = \frac{Δ θ}{Ω}

(6)

where $θ$ indicates the rotor shaft angular position (rad) and Ω denotes the rotor rotational speed (rad s⁻¹).

Under study, Δt was chosen corresponding to 1° rotor rotation per time step.

Numerical models for ZWR performance betterment

The numerical model of ZWR was composed of 4 rotor blades without stator vanes. The performance betterment was based on two principles: the optimization of the ZWR blades number and the addition of stator vanes. To do so, at first, the rotor blades number effect was examined. For that four configurations of ZWR, as highlighted in Figure 8, were putted to test having respectively 4, 5, 7 and 8 rotor blades to find out the optimum configuration.

Figure 8.

Numerical models for the rotor blades number effect: (a) 4 blades, (b) 5 blades, (c) 7 blades, and (d) 8 blades.

Once found, 12 stator vanes as presented in Figure 9 were added to show their effect on the ZWR performances.

Figure 9.

Numerical models for the stator vanes addition effect: (a) without stator and (b) with stator.

Results and discussions

Experimental results

Wind turbine performance characteristics are identified by the static torque coefficient C_Ts, the torque C_T and the power coefficients C_p in addition to the tip speed ratio λ.

These non-dimensional parameters are defined as follows.

C_{Ts} = \frac{T_{s}}{\frac{1}{2} ρ R A {V_{\infty}}^{2}}

(7)

A = D H

(8)

C_{T} = \frac{T}{\frac{1}{2} ρ R A {V_{\infty}}^{2}}

(9)

λ = \frac{R Ω}{V_{\infty}}

(10)

C_{p} = C_{T} λ

(11)

where T_s stands for the static torque (N), V_∞ indicates the incoming wind velocity (m s⁻¹), ρ refers to the air density (kg m⁻³), R corresponds to the rotor radius (m), and A expresses the projected (m²), H the rotor height (m), D is the rotor diameter (m) and Ω expresses the rotor rotational speed (rad s⁻¹).

For the two configurations of the ZWR (without and with stator vanes), the effect of the Reynolds number on the static torque and the torque and the power coefficients are summarized in Figures 10 to 12, respectively.

Figure 10.

Reynolds number effect on ZWR static torque: (a) without stator vanes and (b) with 12 stator vanes.

Figure 11.

Reynolds number effect on ZWR performance without stator vanes: (a) torque coefficient and (b) power coefficient.

Figure 12.

Reynolds number effect on ZWR performance with 12 stator vanes: (a) torque coefficient and (b) power coefficient.

Figure 10 illustrates the variation of the static torque coefficient in terms of incident angle Θ for the examined Reynolds number and for both ZWR configurations. As it is clear, the variation of the static torque coefficient is periodic having a 90° cycle as the rotor possesses four rotating blades. The static torque coefficient increases gradually until reaching its peak and for the rest of the cycle, it declines to attain its minimum at an azimuth around 90°. With the increase in Re, it is clear that the static torque coefficient increases for both rotor configurations. Indeed the self starting ability is more important for higher value of wind velocity subsequently for stronger wind flow (Re = 158,000). Comparing the two ZWR configurations, one can note that the addition of stator vanes affects positively the variation of the static torque. It maximum value reached C_Ts = 0.42 with 12 stator vanes while it was equal to C_Ts = 0.25 without stator vanes. The static torque there by the self starting capacity was also improved with the addition of stator vanes.

From Figures 11(a) and 12(a), the variation of the torque coefficient depending on λ, for both rotor configurations, represents the same variation for the suggested Reynolds numbers. Even with the increase of λ, C_T decays. In effect, more loads are applied, more the ZWR rotation is slowed and consequently the produced torque rises. Besides, it is clear that the curvature corresponding to the strongest flow (Re = 158,000) is above that corresponding to the weakest flows (Re = 143,000 and Re = 111,000). So as the Reynolds number increases, the torque coefficient increases and subsequently the produced torque. Highest values are obtained for Re = 158,000. Moreover, it is to note that with the addition of the stator vanes, the torque coefficient values were improved. Its maximum value passes from C_T,max = 0.09 to C_T,max = 0.15 for the configuration without and with stator vanes, respectively for Re = 158,000.

For both rotor configurations from Figures 11(b) and 12(b), the variation of the power coefficient versus λ indicates the same variation for the suggested Reynolds numbers. It rises with the rise in λ until reaching its highest value (C_p,max) over which it declines despite even the rise in λ. The increase in the Reynolds number subsequently the velocity of the wind flow facing the ZWR, is coupled with the increase in the power coefficient for both configurations. Indeed, the ZWR becomes more efficient at stronger flow. The highest efficiency was obtained for Re = 158,000. Moreover, the addition of the stator vanes affects positively the ZWR performance. In fact, the maximum value of C_p reached C_p,max = 0.074 at λ = 0.99 (Figure 12(b)) while it was equal to C_p,max = 0.032 (Figure 11(b)) for the configuration without stator vanes.

Numerical model validation

A mesh study test was undertaken to ensure that the created mesh adopt for numerical optimization provide relevant and convenient results in good accordance with the experiments. It is to note that, experimental findings used to validate the numerical model are those presented bellow for a ZWR with 4 rotor blades and without stator vanes. To do so, simulations were conducted with five meshes having 74,000, 102,000, 170,000 and 325,000 elements. Table 2 provides a comparison for torque and the power coefficients, for λ = 0.35, between the experimental findings and the numerical ones derived from the meshes under test. The mesh with 74,000 elements is a coarse one and gave the highest value of error (46) relative to the experimental results. It is so unreliable. The mesh with 102,000 elements is a medium one and highlights an error (15%) that is lower than the previous one but still remains important therefore it shows a non-negligible difference over the experimental values. It is also dismissed. The remaining two meshes gave lower error with torque and power coefficients values very close to the experimental ones. As the more the number of mesh elements increase, the more simulation takes time. So the mesh with 170,000 was selected to save time efficiency and as it closely matches the outcomes of the experiments.

Table 2.

Meshing effect on ZWR performance for λ = 0.35.

Elements number	C _T,max		C _p,max		Error (%)
Elements number	Exp	Num	Exp	Num	Error (%)
74,000	0.0714	0.0412	0.025	0.014	46
102,000	0.0714	0.06	0.025	0.021	15
170,000	0.0714	0.067	0.025	0.0236	5.6
325,000	0.0714	0.0704	0.025	0.0247	1.2

For an interval of tip speed ratio values that vary between λ = 0.075 and λ = 0.35, a superposition of experimental and numerical findings is offered in Figure 13. Good agreement was recorded for the tested tip speed ratios. Consequently, the numerical model can be used to forecast the optimization of the ZWR.

Figure 13.

ZWR numerical model validation.

Numerical results

Blades number effect

The rotor blades number effect on the performance of a ZWR without stator vanes was investigated for a Re = 158,000. Numerical torque and power coefficients are highlighted in Figure 14(a) and (b), respectively, for a range of tip speed ratio that vary between λ = 0.075 and λ = 0.22. It is clear that with the increase in the rotor blades number, the torque and the power coefficients of the ZWR were improved. Highest values were recorded with the configuration with 8 blades. Maximum value of the torque coefficient based on Figure 14(a) passes from C_T,max = 0.138 to C_T,max = 0.519 λ = 0.075 for 4 and 8 blades respectively. For the power coefficient and from Figure 14(b), it reached C_p,max = 0.07 with 8 blades while it was C_p,max = 0.0174 with 4 blades for a tip speed ratio λ = 0.22. Indeed the optimum configuration of ZWR derived from the rotor blades number study is that with 8 rotor blades with a significant improvement over the ZWR.

Figure 14.

Rotor blades number effect on ZWR performance: (a) torque coefficient and (b) power coefficient.

Figure 15 shows the velocity fields for the tested configurations with different blades number at λ = 0.22 in the longitudinal plane defined by z = 0 mm. Upstream the ZWR, the velocity is uniform equal to V = 10 m s⁻¹ as established in the boundary conditions. In the region surrounding the ZWR, it is obvious that the velocity decreases slightly. In fact, the rotor plays the role of a barrier ahead the incoming air flow. Maximum velocity values are located at the attack point of the blade above and reached V = 15 m s⁻¹, V = 15.6 m s⁻¹, V = 17 m s⁻¹ and V = 17 m s⁻¹ respectively for 4, the 5, the 7 and the 8-bladed rotors, respectively. This increase in value is untangled by the fact of the creation of a lift force that is perpendicular to the flow direction which causes in part the ZWR rotation. This force is so improved with the increments in rotor blades number and seems more important with the 8-bladed rotor. Besides, flow circulation is clearly noted inside the turbine between the blades which is more notable with the 8-bladed rotor. Downstream the ZWR, a remarkable deceleration is observed for all configurations: A wake phenomenon is introduced. Accordingly, the increase in the number of the blades affects positively the velocity fields and so the wind flow around the ZWR.

Figure 15.

Rotor blades number effect on the velocity fields (m s⁻¹) in the longitudinal plane defined by z = 0 mm: (a) 4 blades, (b) 5 blades, (c) 7 blades, and (d) 8 blades.

Figure 16 shows the distribution of the total pressure for the tested configurations with different blades number at λ = 0.22 in the longitudinal plane defined by z = 0 mm. According to these results, the distribution of total pressure is analogous when comparing the four different configurations: high and low pressure zones are placed on either side of the ZWR. Upstream the ZWR, a notable compression zone is detected, where the total pressure is equal to p = 101,435 Pa, p = 101,444 Pa, p = 101,490 Pa and p = 101,498 Pa respectively for the ZWR with 4 blades, 5 blades, 7 blades and 8 blades. Downstream, a depression zone is highlighted with values around p = 101,354 Pa, p = 101,378 Pa, p = 101,397 Pa and p = 101,414 Pa respectively for the ZWR with 4 blades, 5 blades, 7 blades and 8 blades. This difference in the pressure distribution upstream and downstream the ZWR is the responsible of the rotor rotation and it is clear that is improved with the increase of the blades number and seems more important with 8-bladed rotor. Moreover, the total pressure presents two different features on the concave and the convex part of each blade of the ZWR for the all proposed configurations. with the increase in the blades number, the pressure is improved which improves the force created for the blade rotation. Thus, it is clearly noted that increasing the number of the blades boosted the pressure recovery. This would automatically affect the value of the predicted torque which would be the highest one relative to the 8-blalded ZWR.

Figure 16.

Rotor blades number effect on the total pressure (Pa) distribution in the longitudinal plane defined by z = 0 mm: (a) 4 blades, (b) 5 blades, (c) 7 blades, and (d) 8 blades.

Stator vanes addition effect

In order to further enhance the ZWR performance, 12 stator vanes were attached arround the 8-blade rotor equidistincly as this technique was benifical based on experimental findings. Numerical torque and power coefficients are highlighted in Figure 17(a) and (b), respectively, for a range of tip speed ratio that vary between λ = 0.075 and λ = 0.22. It is obvious that the addition of 12 stator vanes bettered the performance of the 8-bladed ZWR. Based on Figure 17(a), highest value of torque coefficient reached C_T,max = 0.952 at λ = 0.075. As for the power coefficient, the peak attained C_p,max = 0.168 for λ = 0.22. A notable improvement was obtained over the configuration without stator vanes. Thus the optimum configuration of the ZWR is that with 8 rotor blades and with 12 stator vanes.

Figure 17.

Stator vanes addition effect on ZWR performance: (a) torque coefficient and (b) power coefficient.

Figure 18 highlights the distribution of the velocity fields for the 8-blade ZWR with and without stator vanes at λ = 0.22 in the longitudinal plane defined by z = 0 mm. The air stream is uniform at the inlet with a velocity value equal to V = 10 m s⁻¹. Around the ZWR, the wind velocity increases compared to the inlet values and it increases more with the addition of the stator vanes. (V = 13 m s⁻¹ without the stator vanes and for the geometry with stator, it reaches V = 15 m s⁻¹). Maximum value is observed upstream the rotor which is related to the lift force created with the rotor rotation. It is improved with the stator vanes addition and it reaches V = 17 m s⁻¹ while it is equal to V = 15 m s⁻¹ without the stator vanes.

Figure 18.

Stator vanes addition effect on the velocity fields (m s⁻¹) in the longitudinal plane defined by z = 0 mm: (a) without stator vanes and (b) with stator vanes.

Besides, flow circulation is improved inside the turbine between the blades with the stator vanes addition. Downstream the ZWR, a remarkable deceleration is observed: A wake phenomenon is introduced. Accordingly, the addition of stator vanes affects positively the velocity fields and so the wind flows around the ZWR.

Figure 19 shows the distribution of the total pressure for the 8-blade ZWR with and without stator vanes at λ = 0.22 in the longitudinal plane defined by z = 0 mm. High and low pressure zones are highlighted upstream and downstream the ZWR. The pressure difference is responsible of the rotor rotation. With the stator vanes addition the value of the depression zone decreases compared to the configuration without stator vanes. For the compression zone, the pressure value is more important for the configuration with stator vanes. Thus, the pressure recovery is improved with the addition of the stator vanes. Thus, the predicted torque would be improved.

Figure 19.

Stator vanes addition effect on the total pressure (Pa) distribution in the longitudinal plane defined by z = 0 mm: (a) without stator vanes and (b) with stator vanes.

Conclusions

Within this research, a ZWR was investigated experimentally and numerically with the aim of performance betterment. Experimental tests were conducted on a prototype of ZWR made up of 4 rotor blades and 12 removable stator vanes in a wind tunnel for different Reynolds numbers. Highest performances were occurred for Re = 158,000 corresponding to V_∞ = 10 m.s^-1 (C_p,max = 0.074 at λ = 0.99, C_T,max = 0.15 and C_Ts = 0.42). Numerical study was carried out with the use of Ansys Fluent 17.0 software through transient 3D simulations investigating the optimization of the ZWR rotor blades and stator vanes with the aim of performance betterment for Re = 158,000 equivalent to an inlet velocity of V = 10 m s⁻¹. With regard to the rotor blades optimization, for the configuration without stator vanes, it was found that increasing the blades number boosted the performance of the ZWR. Highest power coefficient of C_p,max = 0.07 was obtained for the 8-bladed rotor. The addition of 12 stator vanes improved further the performance and the aerodynamic characteristics of the studied rotor. Maximum power coefficient reached C_p,max = 0.16. The wind flow velocity fields and pressure recovery were also enhanced.

Footnotes

Appendix I

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Abid

Ketata

Lajnef

, et al. (2023) Numerical investigation of greenhouse climate considering external environmental factors and crop position in Sfax central region of Tunisia. Solar Energy 264: 112032.

Alom

Saha

(2019) Influence of blade profiles on Savonius rotor performance: Numerical simulation and experimental validation. Energy Conversion and Management 186: 267–277.

Hassanzadeh

Mohammad Nejad

(2019) Effects of inward and outward overlap ratios on the two-blade Savonius type of vertical axis wind turbine performance. International Journal of Green Energy 16: 1485–1496.

Idrissi

Selmi

Chrigui

(2023) Efficiency improvement of Savonius wind turbine by mean of novel deflector system. Journal of the Brazilian Society of Mechanical Sciences and Engineering 45: 396.

Jaohindy

McTavish

Garde

, et al. (2013) An analysis of the transient forces acting on Savonius rotors with different aspect ratios. Renewable Energy 55: 286–295.

Jeon

Jeong

Pan

, et al. (2015) Effects of the end plates with various shapes and sizes on helical Savonius wind turbines. Renewable Energy 79: 167–176.

Jia

Xia

Zhang

, et al. (2022) Optimal design of Savonius wind turbine blade based on support vector regression surrogate model and modified flower pollination algorithm. Energy Conversion and Management 270: 116247.

Lajnef

Mosbahi

Chouaibi

, et al. (2020) Performance improvement in a helical Savonius wind rotor. Arabian Journal for Science and Engineering 45: 9305–9323.

Layeghmand

Ghiasi Tabari

Zarkesh

(2020) Improving efficiency of Savonius wind turbine by means of an airfoil-shaped deflector. Journal of the Brazilian Society of Mechanical Sciences and Engineering 42: 528.

10.

Mahmoud

El-Haroun

Wahba

, et al. (2012) An experimental study on improvement of Savonius rotor performance. Alexandria Engineering Journal 51: 19–25.

11.

Mohamed

Alqurashi

Thévenin

(2021) Performance enhancement of a Savonius turbine under effect of frontal guiding plates. Energy Reports 7: 6069–6076.

12.

Mosbahi

Derbel

Lajnef

, et al. (2021a) Performance study of twisted Darrieus hydrokinetic turbine with novel blade design. Journal of Energy Resources Technology 143(9): 091302.

13.

Mosbahi

Lajnef

Derbel

, et al. (2021b) Performance improvement of a drag hydrokinetic turbine. Water 13(3): 273.

14.

Patel

Rathod

. (2023) Influence of negative overlap ratio on the performance of semicircular savonius rotor with straight edge extension on overlap region. In: Doolla

Rather

Ramadesigan

(eds) Advances in Clean Energy and Sustainability 2023. ICAER 2022. Green Energy and Technology. Singapore: Springer, pp. 317–330.

15.

Payambarpour

Najafi

Magagnato

(2020) Investigation of deflector geometry and turbine aspect ratio effect on 3D modified in-pipe hydro Savonius turbine: Parametric study. Renewable Energy 148: 44–59.

16.

Putri

Yuwono

Rustam

, et al. (2019) Experimental studies on the effect of obstacle upstream of a Savonius wind turbine. SN Applied Sciences 1: 1216.

17.

Saikot

MMH

Rahman

Hosen

, et al. (2023) Savonius wind turbine performance comparison with one and two porous deflectors: A CFD study. Flow Turbulence Combust 111: 1–25.

18.

Shamsuddin

MSM

Kamaruddin

(2023) Experimental study on the characterization of the self-starting capability of a single and double-stage Savonius turbin. Results in Engineering 17: 100854.

19.

Sharma

(2016) Performance improvement of Savonius rotor using multiple quarter blades: A CFD investigation. Energy Conversion and Management 127: 43–54.

20.

Shende

Patidar

Baredar

, et al. (2022) A review on comparative study of Savonius wind turbine rotor performance parameters. Environmental Science and Pollution 29: 69176–69196.