the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Difference in load predictions obtained with effective turbulence vs. a dynamic wake meandering modeling approach
Kelsey Shaler
Jason Jonkman
Abstract. According to the international standard for wind turbine design, the effects of wind turbine wakes on structural loads can be considered in two ways: (1) by augmenting the ambient turbulence levels with the effective turbulence model (EFF) and then calculating the resulting loads and (2) by performing dynamic wake meandering (DWM) simulations, which compute wake effects and loads for all turbines in a farm at once. There is no definitive answer in scientific literature as to the consequences of choosing one model over the other, but the two approaches are unarguably very different. The work presented here expounds on these differences and investigates to what extent they affect the simulated structural loads. We consider an idealized 4x4 rectangular array of National Renewable Energy Laboratory 5 MW wind turbines with a spacing of 5 by 8 rotor diameters, and three wind speed scenarios at high ambient turbulence. Load simulations are performed in OpenFAST with EFF and in FAST.Farm with the DWM implementation. We compare ambient turbulence, wind farm turbulence, and loads between both approaches. When omnidirectional results are compared, EFF wind farm turbulence intensity is consistently higher by 0.2 % (above rated wind speed) to 2.7 % (below rated wind speed). However, for certain wind directions, the EFF turbulence can be lower than FAST.Farm by almost 9 %. Wind speeds within the farm were found to differ by up to 3 m s-1 due to the lack of wake deficits in the EFF approach, leading to longer tails toward low values in the FAST.Farm mean load distributions. Consistent with the turbulence results, the median EFF load standard deviations are also consistently higher, by a maximum of 20 % and 17 % for blade-root out-of-plane and tower base fore-aft moments, respectively.
- Preprint
(18544 KB) - Metadata XML
- BibTeX
- EndNote
Paula Doubrawa et al.
Status: closed
-
AC1: 'Comment on wes-2023-26', Paula Doubrawa, 11 Apr 2023
There is a typo on Equation 3. An exponent of 2 is missing around the entire denominator of the fraction. Compare to equation for \hat{sigma_t} on page 112 (annex E) of IEC 61400-1:2019. It is only a typo. The code used in the simulations is correct. The typo will be corrected when the manuscript is peer reviewed and resubmitted.
Thank you Shadan Mozafari from DTU for catching it.
Citation: https://doi.org/10.5194/wes-2023-26-AC1 -
RC1: 'Comment on wes-2023-26', Anonymous Referee #1, 21 Apr 2023
General comment
In this article, the effective turbulence approach (EFF) and the dynamic wake meandering (DWM) approach are compared in terms of freestream turbulence, wind farm turbulence, load standard deviations, spectra, and means on a 4x4 rectangular array of NREL 5 MW wind turbines.
- The article is well written and clear.
- Each step of the two approaches is well detailed by the authors, what makes the article clear for the reader not acquainted with those approaches.
- The interest of the study is clearly stated in the introduction, i.e. to allow wind farm modelling experts to be informed and intentional when choosing simulation tools and making design decisions, and a position is taken by the authors in the conclusion.
This article has certainly a value for the wind energy research community, and I suggest the manuscript for publication after addressing the following comments.
Specific comments
L88: “CT is the wind turbine thrust coefficient” --> CT is the neighboring (?) wind turbine thrust coefficient
L102: “Note that wake-added turbulence is a forthcoming capability of FAST.Farm that was not available in the model version used here. That said, the ambient turbulence intensities simulated in the wind scenarios are high enough that the absence of wake-added turbulence would not likely impact the conclusions of this study (Shaler and Jonkman, 2021).”
To the best of my understanding, the EFF approach is based on the computation of an effective turbulence to consider the influence of the adjacent wind turbines on the target turbine, i.e. to consider the influence of the wake-added turbulence. However, regarding the FAST.Farm computations, you justify that the wake-added turbulence would not impact the conclusions because of the high turbulence wind scenarios. Since the wind scenarios are similar between EEF and FAST.Farm, I see a contradiction there. Could you comment on this?
L210: I suggest adding a subsection regarding the structural loading computation inside the section "Methods" and refer to steps FF.3 and EFF.3 of Fig. 5.
L245: In Figure 10, some turbulence intensities obtained with FAST.Farm (directions 0° -90°) for non-waked turbines are much higher than the average freestream turbulence, e.g. T1 for 30° in Figure 10 (c). Considering T1, it is also the case to a lesser extent for other wind directions in Figure 10 (a) and 10 (b). Could you comment on this?
L332: Same remark as before regarding the effect of the wake-added turbulence.
L335: “On average, our results agree with previous studies that compared the EFF to measurements (Argyle et al., 2018; Reinwardt et al., 2018) and DWM predictions (Reinwardt et al., 2018) and found EFF to overestimate turbulence levels.”
If found this sentence a bit in contradiction with one sentence of the introduction (L60): “The work that we present here was motivated by the small number of published studies on this topic and the lack of consistency among them.” I therefore suggest slightly rephrasing to insist on the fact that your results agree with the specific works of Argyle et al., 2018 and Reinwardt et al.,2018 instead of using “agree with previous studies”.
Minor comments
L33: “helps answer” --> helps to answer
L150: “andtakenas” --> typo
L284: In the legend of Figure 14, I suggest keeping using EFF instead of ETM for a better consistency within the article.
L315: Same remark as for Figure 14.
L373: “to support adoption” --> to support the adoption
Citation: https://doi.org/10.5194/wes-2023-26-RC1 - AC2: 'Reply on RC1', Paula Doubrawa, 19 Jul 2023
-
RC2: 'Comment on wes-2023-26', Anonymous Referee #2, 27 Jun 2023
General comments
This article is well-structured and addresses the relevant topic of quantifying the difference in structural fatigue loading between the dynamic wake meandering model and the effective turbulence. All main results are based on simulation in accordance with the IEC design standards for a 4x4 rectangular grid layout using the NREL 5MW reference turbine.
• The introduction presents previous related work and clearly states the intent and scientific relevance and contribution of this paper.
• The methods and simulations provide an overview of the analysis scope and differences between the two turbulence models.
• The results are clearly presented and supported by relevant illustrations.
• The conclusion effectively compares the results and relates them to the stated intent in the introduction. It also provides a clear recommendation for the wind energy community and suggests further research directions.
The topic in this paper is highly relevant to the wind energy industry and has the potential to significantly enhance decision-making in structural load assessment. I recommend addressing the following comments before publishing.
Specific commentsEq3: It would be helpful if C_T is expressed as a function of V_hub.
L 88: C_T is the thrust coefficient of neighbouring turbines.
L 103-105: The statement claiming that the absence of wake-added turbulence would not impact conclusions is quite bold. One of the main results of this paper is that the effective turbulence model results in higher turbulence levels than the dynamic wake meandering model, and at least some of this difference can be attributed to the missing wake-added turbulence. Please provide further insight on this matter.
L 137-139: It is not clear if the stated turbulence intensities are characteristic values.
L 150: The end of bullet one requires editing since the text is in italics when it is not supposed to be.
L 155-157: The choice of using 80% directly influences your results. Please explain how this choice impacts your conclusions, and if it is considered insignificant, provide an argument as to why.
L 205-206: The numbers stated in the text differ slightly from those shown in Figure 8.
L 284-285: Load standard deviations are introduced rapidly. Please provide a more detailed explanation.
L 290-293: Fatigue loads are heavily influenced by the highest load cycles (due to each load cycle being raised to the power of "m" when calculating its contribution to fatigue damage). It would be interesting to include a comparison of higher-order raw moments of the load standard deviation distribution as a supplement to comparing medians.
L 342-345: Similar to the previous comment, please comment on the potential effect of narrow versus wide distributions.
L 365-372: The industry is moving towards estimating fatigue loads by considering the entire ambient turbulence distribution rather than relying on the characteristic turbulence (i.e., integrate fatigue loads across the ambient turbulence distribution for each wind speed). This is intractable to do via aero-elastic simulation and therefore surrogate models are being developed. Such surrogate models are relatively easy
to train for the effective turbulence as it does not require a lot of parameters – as opposed to DWM. It would strengthen the paper to briefly discuss this potential issue of integrating the DWM model into the current practice of wind farm design.Citation: https://doi.org/10.5194/wes-2023-26-RC2 - AC3: 'Reply on RC2', Paula Doubrawa, 21 Jul 2023
Status: closed
-
AC1: 'Comment on wes-2023-26', Paula Doubrawa, 11 Apr 2023
There is a typo on Equation 3. An exponent of 2 is missing around the entire denominator of the fraction. Compare to equation for \hat{sigma_t} on page 112 (annex E) of IEC 61400-1:2019. It is only a typo. The code used in the simulations is correct. The typo will be corrected when the manuscript is peer reviewed and resubmitted.
Thank you Shadan Mozafari from DTU for catching it.
Citation: https://doi.org/10.5194/wes-2023-26-AC1 -
RC1: 'Comment on wes-2023-26', Anonymous Referee #1, 21 Apr 2023
General comment
In this article, the effective turbulence approach (EFF) and the dynamic wake meandering (DWM) approach are compared in terms of freestream turbulence, wind farm turbulence, load standard deviations, spectra, and means on a 4x4 rectangular array of NREL 5 MW wind turbines.
- The article is well written and clear.
- Each step of the two approaches is well detailed by the authors, what makes the article clear for the reader not acquainted with those approaches.
- The interest of the study is clearly stated in the introduction, i.e. to allow wind farm modelling experts to be informed and intentional when choosing simulation tools and making design decisions, and a position is taken by the authors in the conclusion.
This article has certainly a value for the wind energy research community, and I suggest the manuscript for publication after addressing the following comments.
Specific comments
L88: “CT is the wind turbine thrust coefficient” --> CT is the neighboring (?) wind turbine thrust coefficient
L102: “Note that wake-added turbulence is a forthcoming capability of FAST.Farm that was not available in the model version used here. That said, the ambient turbulence intensities simulated in the wind scenarios are high enough that the absence of wake-added turbulence would not likely impact the conclusions of this study (Shaler and Jonkman, 2021).”
To the best of my understanding, the EFF approach is based on the computation of an effective turbulence to consider the influence of the adjacent wind turbines on the target turbine, i.e. to consider the influence of the wake-added turbulence. However, regarding the FAST.Farm computations, you justify that the wake-added turbulence would not impact the conclusions because of the high turbulence wind scenarios. Since the wind scenarios are similar between EEF and FAST.Farm, I see a contradiction there. Could you comment on this?
L210: I suggest adding a subsection regarding the structural loading computation inside the section "Methods" and refer to steps FF.3 and EFF.3 of Fig. 5.
L245: In Figure 10, some turbulence intensities obtained with FAST.Farm (directions 0° -90°) for non-waked turbines are much higher than the average freestream turbulence, e.g. T1 for 30° in Figure 10 (c). Considering T1, it is also the case to a lesser extent for other wind directions in Figure 10 (a) and 10 (b). Could you comment on this?
L332: Same remark as before regarding the effect of the wake-added turbulence.
L335: “On average, our results agree with previous studies that compared the EFF to measurements (Argyle et al., 2018; Reinwardt et al., 2018) and DWM predictions (Reinwardt et al., 2018) and found EFF to overestimate turbulence levels.”
If found this sentence a bit in contradiction with one sentence of the introduction (L60): “The work that we present here was motivated by the small number of published studies on this topic and the lack of consistency among them.” I therefore suggest slightly rephrasing to insist on the fact that your results agree with the specific works of Argyle et al., 2018 and Reinwardt et al.,2018 instead of using “agree with previous studies”.
Minor comments
L33: “helps answer” --> helps to answer
L150: “andtakenas” --> typo
L284: In the legend of Figure 14, I suggest keeping using EFF instead of ETM for a better consistency within the article.
L315: Same remark as for Figure 14.
L373: “to support adoption” --> to support the adoption
Citation: https://doi.org/10.5194/wes-2023-26-RC1 - AC2: 'Reply on RC1', Paula Doubrawa, 19 Jul 2023
-
RC2: 'Comment on wes-2023-26', Anonymous Referee #2, 27 Jun 2023
General comments
This article is well-structured and addresses the relevant topic of quantifying the difference in structural fatigue loading between the dynamic wake meandering model and the effective turbulence. All main results are based on simulation in accordance with the IEC design standards for a 4x4 rectangular grid layout using the NREL 5MW reference turbine.
• The introduction presents previous related work and clearly states the intent and scientific relevance and contribution of this paper.
• The methods and simulations provide an overview of the analysis scope and differences between the two turbulence models.
• The results are clearly presented and supported by relevant illustrations.
• The conclusion effectively compares the results and relates them to the stated intent in the introduction. It also provides a clear recommendation for the wind energy community and suggests further research directions.
The topic in this paper is highly relevant to the wind energy industry and has the potential to significantly enhance decision-making in structural load assessment. I recommend addressing the following comments before publishing.
Specific commentsEq3: It would be helpful if C_T is expressed as a function of V_hub.
L 88: C_T is the thrust coefficient of neighbouring turbines.
L 103-105: The statement claiming that the absence of wake-added turbulence would not impact conclusions is quite bold. One of the main results of this paper is that the effective turbulence model results in higher turbulence levels than the dynamic wake meandering model, and at least some of this difference can be attributed to the missing wake-added turbulence. Please provide further insight on this matter.
L 137-139: It is not clear if the stated turbulence intensities are characteristic values.
L 150: The end of bullet one requires editing since the text is in italics when it is not supposed to be.
L 155-157: The choice of using 80% directly influences your results. Please explain how this choice impacts your conclusions, and if it is considered insignificant, provide an argument as to why.
L 205-206: The numbers stated in the text differ slightly from those shown in Figure 8.
L 284-285: Load standard deviations are introduced rapidly. Please provide a more detailed explanation.
L 290-293: Fatigue loads are heavily influenced by the highest load cycles (due to each load cycle being raised to the power of "m" when calculating its contribution to fatigue damage). It would be interesting to include a comparison of higher-order raw moments of the load standard deviation distribution as a supplement to comparing medians.
L 342-345: Similar to the previous comment, please comment on the potential effect of narrow versus wide distributions.
L 365-372: The industry is moving towards estimating fatigue loads by considering the entire ambient turbulence distribution rather than relying on the characteristic turbulence (i.e., integrate fatigue loads across the ambient turbulence distribution for each wind speed). This is intractable to do via aero-elastic simulation and therefore surrogate models are being developed. Such surrogate models are relatively easy
to train for the effective turbulence as it does not require a lot of parameters – as opposed to DWM. It would strengthen the paper to briefly discuss this potential issue of integrating the DWM model into the current practice of wind farm design.Citation: https://doi.org/10.5194/wes-2023-26-RC2 - AC3: 'Reply on RC2', Paula Doubrawa, 21 Jul 2023
Paula Doubrawa et al.
Paula Doubrawa et al.
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
767 | 169 | 20 | 956 | 9 | 8 |
- HTML: 767
- PDF: 169
- XML: 20
- Total: 956
- BibTeX: 9
- EndNote: 8
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1