Study of Aerodynamic Characteristics of Asymmetrical Blades and a Wind-Driven Power Plant with a Vertical Axis of Rotation
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsMajor issues:
1) No quantitative comparative analysis between numerical calculations and wind tunnel test data were provided, Why? Is there much deviation in the quantitative results obtained from the present numerical simulation?
2) There is a little information provided for numerical simulation methods and computational settings. Is the 10 layer boundary layer grid sufficient? What is the Yplus value of the first layer grid and what is the normal growth rate of the boundary layer grid?
Author Response
Comments 1: No quantitative comparative analysis between numerical calculations and wind tunnel test data were provided, Why? Is there much deviation in the quantitative results obtained from the present numerical simulation?
Response 1: Thank you for pointing out this omission. We have supplemented the article with graphs (Figures 18 and 19) and conducted a comparative analysis of numerical calculation and experimental data. The results of numerical simulation demonstrate a high degree of agreement with experimental data. Pressure versus velocity graphs show a discrepancy of no more than 5%. This agreement confirms the reliability of the model and its ability to correctly describe the studied physical processes.
Comments 2: There is a little information provided for numerical simulation methods and computational settings. Is the 10 layer boundary layer grid sufficient? What is the Yplus value of the first layer grid and what is the normal growth rate of the boundary layer grid?
Response 2: For models such as SST k-ω, 10 layers may be sufficient if Y+ of the first layer is within <1. If the flow around the object is complex (for example, strong velocity gradients, flowing around sharp corners), a larger number of layers (15-20) may be required. In this work, a grid of 15 layers of triangular cells was used. The value of the Y+ layer was estimated in the range of 0.01–0.020, which corresponds to Y+ in the recommended range. The growth of the elements in the normal direction is neat and gradual. The estimated value of the normal growth rate is 1.2–1.3 (or 20-30% increase in height between adjacent layers).
Reviewer 2 Report
Comments and Suggestions for AuthorsIn this article, the authors discuss important and current issues related to increasing the efficiency of obtaining energy from wind using VAWT with asymmetric blades.
The paper analyzes the properties of the asymmetric profile in terms of its use in a vertical-axis turbine. The paper contains the results of laboratory-scale tests and numerical simulations.
After a broad and thorough analysis of literature data, the authors present the results of their own research carried out at a laboratory stand.
Questions and doubts:
1. Regarding the detailed measurement methodology, we refer to the literature that is not available online, preventing a full assessment of the measurements.
2. Additionally, the literature review does not contain references to the latest publications. Most of them are references to publications that were published over 10 years ago.
3. Most of the presented graphs have axes that do not intersect at zero, which, in my opinion, is a significant error. If we show a relationship where the scales on the axes start from "0" and the axes do not intersect at "0" - in my opinion, the recipient is misled.
4. Figure 9 does not describe the axis and is also on a scale different from the previously presented graph.
5. Figures 10 I 12 - It is unclear to me what is presented on the graph's Y axis and what the "H" unit is, whether it is an SI unit. In addition, the X-axis only has a unit given.
6. Does graph 11 present the results for an unloaded turbine?
7. The presented information on numerical simulations does not contain information on the quality of the mesh or whether verification of the independence of the solution from the mesh was performed. Analyzing the figure showing the mesh, it can be concluded that a boundary layer was not used, which will probably be important considering the analyzed case. In addition, did the authors verify the effect of the size of the computational domain on the result? From the presented results, it can be seen that on the domain's boundary, the analyzed values ​​have a variable character - and they should not have it. Additionally, the authors do not provide full information on the boundary conditions, turbulence model, etc. - information of this type should be supplemented.
8. The legend on the graphs in Fig. 15 and 16 is hardly legible. In addition, the drawings and component results are presented in different scales. Since the simulation was carried out for the exact value of the inflow velocity, the obtained result should have similar values ​​for both the pressure distribution and the velocity. The discrepancies that we see, for example, in the vector velocity distributions have differences from 3.7 m/s to 22.7 m/s of the maximum velocity value. If the inflow velocity is 3 m/s and the maximum value is 22.7 m/s - what is the reason for such a large maximum value, such a significant difference would require a broader justification.
9. A very important question that arises when reading the article is why the authors do not compare the results of the laboratory tests with the simulation tests in any way. Having a result from a laboratory test imposes the validation of the numerical model - there is no information on this in the article. Such a comparison would be very valuable information, and would also constitute an excellent verification of the simulation results.
Author Response
Comments 1: Regarding the detailed measurement methodology, we refer to the literature that is not available online, preventing a full assessment of the measurements:
Response 1: Thanks for the comment. We have added available up-to-date links that meet current standards.
Comments 2: Additionally, the literature review does not contain references to the latest publications. Most of them are references to publications that were published over 10 years ago:
Response 2: We have taken this observation into account and updated the literature to include more relevant publications of recent years.
Comments 3: Most of the presented graphs have axes that do not intersect at zero, which, in my opinion, is a significant error. If we show a relationship where the scales on the axes start from "0" and the axes do not intersect at "0" - in my opinion, the recipient is misled:
Response 3: We recognize that graphs with axes that do not intersect at zero can be misleading. In the article, all graphs are corrected in such a way as to ensure the correct display of dependencies and avoid data distortion.
Comments 4: Figure 9 does not describe the axis and is also on a scale different from the previously presented graph:
Response 4: Figure 9 has been brought into line with other figures.
Comments 5: Figures 10 I 12 - It is unclear to me what is presented on the graph's Y axis and what the "H" unit is, whether it is an SI unit. In addition, the X-axis only has a unit given:
Response 5: Thank you for pointing out the ambiguities. The Y-axis represents the value of the thrust force, measured in Newtons. A description of the velocity value has been added along the X-axis to eliminate uncertainty.
Comments 6: Does graph 11 present the results for an unloaded turbine?
Response 6: Graph 11 really represents the results for an unloaded turbine, since the generator was not connected to relieve the load.
Comments 7: The presented information on numerical simulations does not contain information on the quality of the mesh or whether verification of the independence of the solution from the mesh was performed. Analyzing the figure showing the mesh, it can be concluded that a boundary layer was not used, which will probably be important considering the analyzed case. In addition, did the authors verify the effect of the size of the computational domain on the result? From the presented results, it can be seen that on the domain's boundary, the analyzed values ​​have a variable character - and they should not have it. Additionally, the authors do not provide full information on the boundary conditions, turbulence model, etc. - information of this type should be supplemented:
Response 7: The quality of the calculated grid and the verification of the independence of the solution from the grid: The article states that the calculation grid was created in the ANSYS Meshing environment. The total volume of the calculated grid was 159,130 cells. 15 layers of triangular cells were used to model the boundary layer, which ensures sufficient accuracy in key areas. The verification of the independence of the solution from the grid was carried out by changing the density of the cells, and the results confirmed that variations in the grid do not significantly affect the results. A comparison of numerical data with experimental data showed a discrepancy of less than 5%, which indicates a high accuracy of the model.
The graph shows a study of grid independence for the maximum number of cells N=159130. It can be seen that with an increase in the number of cells, the values of the drag coefficient Cx stabilize, and at N=150000 and N=159130 there are practically no discrepancies (Cx=1.38). This confirms that the selected number of cells ensures sufficient accuracy of calculations.
Using the boundary layer: The boundary layer was modeled using 15 layers of cells with a minimum thickness sufficient to correctly account for changes in velocity and pressure. This is confirmed by contour graphs of pressure and velocity distribution (Figures 15 and 16), which demonstrate the accuracy of numerical modeling in areas of high gradient.
Influence of the size of the design area: The size of the design area was chosen based on the geometry of the experimental installation (the size of the working part of the wind tunnel: 0.5 m × 0.8 m). The control of the stability of the values at the boundaries of the calculated area showed the absence of a significant influence of the boundaries on the studied area, which is confirmed by the stability of the pressure and velocity distributions at the boundaries.
K-ω SST was used as a turbulence model, which is specially adapted for problems with high gradients and is applicable for calculations of external flow. The following boundary conditions were established: the input current with a constant velocity of 3 m/s, the angles of attack varied from 0° to 180° in increments of 30°.
The presented graphs of numerical data (Figures 15 and 16) show the stability of the values at the boundaries of the region, which confirms the correct choice of the size of the calculated area and grid. The variability of the values at the boundary is explained by the interaction of the flow with the object under study, which is natural for problems with high turbulence.
The use of the k-ω SST model made it possible to take into account the influence of turbulent phenomena on the distribution of pressure and velocity, especially in the boundary layer region. This is confirmed by comparing experimental and numerical data, which showed a good correspondence, which indicates the reliability and accuracy of the calculations performed.
Comments 8: The legend on the graphs in Fig. 15 and 16 is hardly legible. In addition, the drawings and component results are presented in different scales. Since the simulation was carried out for the exact value of the inflow velocity, the obtained result should have similar values ​​for both the pressure distribution and the velocity. The discrepancies that we see, for example, in the vector velocity distributions have differences from 3.7 m/s to 22.7 m/s of the maximum velocity value. If the inflow velocity is 3 m/s and the maximum value is 22.7 m/s - what is the reason for such a large maximum value, such a significant difference would require a broader justification:
Response 8: The legends on graphs 15 and 16 have been redesigned to improve readability. The scales are unified to make the results more visible.
Comments 9: A very important question that arises when reading the article is why the authors do not compare the results of the laboratory tests with the simulation tests in any way. Having a result from a laboratory test imposes the validation of the numerical model - there is no information on this in the article. Such a comparison would be very valuable information, and would also constitute an excellent verification of the simulation results:
Response 9: We have supplemented the article with graphs (Figures 17 and 18) and conducted a comparative analysis of numerical calculation and experimental data. The results of numerical simulation demonstrate a high degree of agreement with experimental data. Pressure versus velocity graphs show a similarity not exceeding 5%. This agreement confirms the reliability of the model and its ability to correctly describe the studied physical processes.
Author Response File: 
Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors studied lift and drag coefficients at various angles of attack and flow speeds of an asymmetrical balde in wind tunnel experiments, alongside numerical simulations conducted in ANSYS.
The numerical simulation section is not well presented, and the studies were made separately, and no relation was made between experimental and numerical results to validate one of them. The author studied only one blade not three-bladed Darrieus rotors, as it was mentioned in the abstract. And more analyses should be added to the paper to improve its qualtiy/novelty and scientific soundness.
Also there are significant similarities between the papers below:
"Investigation of self-starting and high rotor solidity on the performance of a three S1210 blade H-type Darrieus rotor"
"Studies of some high solidity symmetrical and unsymmetrical blade H-Darrieus rotors with respect to starting characteristics, dynamic performances and flow physics in low wind streams"
Author Response
Comments 1: The numerical simulation section is not well presented, and the studies were made separately, and no relation was made between experimental and numerical results to validate one of them. The author studied only one blade not three-bladed Darrieus rotors, as it was mentioned in the abstract. And more analyses should be added to the paper to improve its qualtiy/novelty and scientific soundness.
Response 1: Thank you for pointing out this omission. When creating a three-bladed structure, the data of one blade is scaled, since the blades are the same in geometry. The models used to analyze a single blade are also valid for a system of three blades, as shown in the layout. The study of one blade is sufficient to understand its aerodynamic characteristics. This data can be applied to the design of rotors with multiple blades, since the aerodynamic parameters of individual blades are scalable and applicable to the complete design.
Comments 2: Also there are significant similarities between the papers below:
"Investigation of self-starting and high rotor solidity on the performance of a three S1210 blade H-type Darrieus rotor"
"Studies of some high solidity symmetrical and unsymmetrical blade H-Darrieus rotors with respect to starting characteristics, dynamic performances and flow physics in low wind streams"
Response 2: These articles have been studied and added to the article.
Round 2
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors say that the comparative analysis of the experimental data (Figures 6 and 7) and the numerical calculation data (Figures 18 and 19) reveals a close match between the numerical simulation and experimental results. The numerical model correctly reflects the patterns and dynamics of the process. This is confirmed by the proximity of the simulation and experimental results. Analysis of the data shows that the relative error does not exceed 5%.
The author must clarify the given 5% error by providing a table of a detailed comparison of the numerical and experimental drag and lift coefficients.
Also figures 6 and 18 and figures 7 and 19 should be superposed to clearly see the difference between the given results.
Author Response
Comments 1: The author must clarify the given 5% error by providing a table of a detailed comparison of the numerical and experimental drag and lift coefficients.
Response 1: Thank you for the comment. We agree that clarification of the 5% error requires additional explanation. For this purpose, the article has been supplemented with Tables 1 and 2, which provide a detailed comparison of numerical and experimental values of drag and lift coefficients for each of the studied angles of attack and flow velocities.
Comments 2: Also figures 6 and 18 and figures 7 and 19 should be superposed to clearly see the difference between the given results.
Response 2: Thank you for the suggestion. We agree that overlaying Figures 6 and 18, as well as Figures 7 and 19, would clearly demonstrate the differences between the results of numerical modeling and experimental investigations. In the revised article, combined graphs have been presented, making the comparison more comprehensible and visually clear.
Author Response File: Author Response.pdf
