Preprints
https://doi.org/10.5194/wes-2021-55
© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/wes-2021-55
© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
15 Jun 2021
| 15 Jun 2021
Status: this preprint has been withdrawn by the authors.
Wind Turbine Ice Detection Using AEP Loss Method – A Case Study
Abstract. Ice detection of wind turbine and estimating the resultant production losses is challenging, but very important, as wind energy project decisions in cold regions are based on these estimated results. This paper describes the comparison of a statistical (T19IceLossMethod) and numerical (Computational Fluid Dynamics, CFD) case study of wind resource assessment and estimation of resultant Annual Energy Production (AEP) due to ice of a wind park in ice prone cold region. Three years Supervisory Control and Data Acquisition (SCADA) data from a wind park located in arctic region is used for this study. Statistical analysis shows that the relative power loss due to icing related stops is the main issue for this wind park. To better understand the wind flow physics and estimation of the wind turbine wake losses, Larsen wake model is used for the CFD simulations, where results show that it is important to use the wake loss model for CFD simulations of wind resource assessment and AEP estimation of a wind park. A preliminary case study about wind park layout optimization has also been carried out which shows that AEP can be improved by optimizing the wind park layout and CFD simulations can be a good tool.
This preprint has been withdrawn.
How to cite. Jin, J. Y., Karlsson, T., and Virk, M. S.: Wind Turbine Ice Detection Using AEP Loss Method – A Case Study, Wind Energ. Sci. Discuss. [preprint], https://doi.org/10.5194/wes-2021-55, 2021.
Download & links
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(1764 KB)
Interactive discussion
Status: closed
Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor
| : Report abuse
-
RC1: 'Comment on wes-2021-55', Anonymous Referee #1, 13 Jul 2021
The manuscript deals with a comparison between two approaches for estimating the icing effect: a statistical method and CFD.
Although the main idea of comparing such approaches together is interesting, the whole paper seems vague to me.- The goal of the study and the research question is not clearly stated, not in abstract, nor in conclusion and not even in the discussions. It is confusing why the comparison is implemented whether for validating CFD or validating the statistical method? Which method is faster or more efficient to be implemented comparing the accuracy and the computational costs?
- The results presentation needs to be improved via further postprocessing. They look more like the raw data in the tables. The only clear message of the paper to me is that the wake loss model is improving the CFD simulation. This is not sufficient for a comparison research. Perhaps the authors have gained more conclusions but the results are not well-organized or classified.
- The text is not smooth to read in addition to many grammatically issues and a vague structure.
- My detailed comments are uploaded in the pdf file, where I addressed the specific parts individually. Hope it helps to improve the paper and I believe that the idea of the paper is potential to be developed, but at this stage the research is not prepared to be published. I do not recommend it for publication.
-
AC1: 'Reply on RC1', JIAYI JIN, 31 Jul 2021
The comment was uploaded in the form of a supplement: https://wes.copernicus.org/preprints/wes-2021-55/wes-2021-55-AC1-supplement.zip
-
RC2: 'Comment on wes-2021-55', Anonymous Referee #2, 04 Oct 2021
This paper used two approaches to estimate the AEP loss of a wind park in Norway due to icing.The first approach used a statistical investigation of the SCADA data, while the second approach compared the observed AEP with the results from a CFD study. The paper also tried to optimize the wind farm using the CFD tool, but this didn't seem to take into account the impact of icing, so it wasn't clear how that related to the main theme of the paper.
The main goal of the paper, to evaluate different approaches to estimating AEP loss is interesting, but the paper was quite challenging to follow and it isn't really clear to me what the main outcome of the study was.
Major comments
- The paper as a whole was challenging to read, there were many typos, and the grammer was not very good. The structure also was not laid out for the reader making it hard to follow what the recipe was to get to the end resuts.
- Several topics were only vaguely discussed and were difficult to understand. For example, there were several places where it was stated that the T19Method would be compared to the SCADA data, but the T19Method uses the SCADA data to get its results. I kept expecting there to be an ice detector or some other method to determine the icing periods or loss, but they weren't found.
- None of the methods are clearly described in such a way to allow the reader to be able to replicate the study. The T19IceLossMethod is probably the best described as it is the simplest method, but it is not totally clear how stops were detected, or if duration was used at all for determining icing periods. The Numerical model section (3.2) does not describe any of the input data, what was used to build the 3D terrain model, where did the roughness data come from, what was used for the inflow wind profile? In section 2, it is mentioned that a Weibull fit is carried out, but nothing about the method for fitting was described. While in all subsections in section 4 there were various methods applied that aren't described (4.1: how were the values of 57.1, 54.6, and 55.5% reached?; What is the % loss of total hour colume in table 6?; 4.2: Where can we see the vortex motion in figure 6 that you describe in this section?; 4.3 What is the T19IceLossMethod using in table and figure 7, so far all descriptions of that method produce results as % of ice loss, how do you then get an AEP?
- The data in the tables and figures was often hard to interpret and the reader was not guided to help understand it, rather than having discussion of the results, they were often just summarize in the text.
Minor comments:
Here I will mostly focus on the figures, tables, and references- Shouldn't the reference to the cold climate market either be Timo's presentation at winterwind 2017 (https://winterwind.se/wp-content/uploads/2015/08/9_1_24_Karlsson_IEA_Task_19_-_Cold_climate_wind_power_market_study_2015-2020_Pub_v1.pdf), or the Wind Power Monthly article (https://www.windpowermonthly.com/article/1403504/emerging-cold)?
- On line 38, you start of the paragraph as Gravdahl et al (Davis et al., 2016) it is unclear what the reference you were going for was, and the Gravdahl reference doesn't show up in the references.
- Table 1: it was unclear to me at first what was being show. I am now pretty confident that it contains the mean values for the 14 turbines, but it isn't clear what the reader should get from this table.
- Table 4: perhaps this information could be better portrayed through histograms of the different cell sizes, for example (https://link.springer.com/article/10.1007/s10546-020-00591-0/figures/3)
- Table 5: Is this the percentage of AEP loss? Something else?
- Table 6: If table 5 is correct, was there really 29.6% loss of AEP at turbine 13 with only 19.6 hours of down time, that seems quite extreme.
- Table 6: As mentioned above, there doesn't seem to be any description of that the % loss of total hour is, it seems like it might be the sum of the other columns by year, but should c be subtracted, and not all of the values sum correctly so I am a bit lost.
- Figure 5 is really hard to review, also it is known that there are a lot of points in certain parts of the power curve, consider instead using a hexbin plot like done in Figure 2 from (https://onlinelibrary.wiley.com/doi/10.1002/we.1878), for specific parts of the power curve you want to highlight. I was surprised to not see this paper referenced as it seems to be one of the closer papers to your work.
- Table 7: This seems a bit conterintuitive to me perhaps because in table 5 you were showing the percent loss, while here are showing the percent of the potential. It might be easier for the reader to keep using the same approach to describing the loss.
- Figure 7: What should be a better CFD result? I guess the closer the value is to the T19LossMethod? Since the SCADA data include icing effects that don't model in your CFD results.
- Figure 7: What does this figure add that you don't get from the table above?
Citation: https://doi.org/10.5194/wes-2021-55-RC2
Interactive discussion
Status: closed
Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor
| : Report abuse
-
RC1: 'Comment on wes-2021-55', Anonymous Referee #1, 13 Jul 2021
The manuscript deals with a comparison between two approaches for estimating the icing effect: a statistical method and CFD.
Although the main idea of comparing such approaches together is interesting, the whole paper seems vague to me.- The goal of the study and the research question is not clearly stated, not in abstract, nor in conclusion and not even in the discussions. It is confusing why the comparison is implemented whether for validating CFD or validating the statistical method? Which method is faster or more efficient to be implemented comparing the accuracy and the computational costs?
- The results presentation needs to be improved via further postprocessing. They look more like the raw data in the tables. The only clear message of the paper to me is that the wake loss model is improving the CFD simulation. This is not sufficient for a comparison research. Perhaps the authors have gained more conclusions but the results are not well-organized or classified.
- The text is not smooth to read in addition to many grammatically issues and a vague structure.
- My detailed comments are uploaded in the pdf file, where I addressed the specific parts individually. Hope it helps to improve the paper and I believe that the idea of the paper is potential to be developed, but at this stage the research is not prepared to be published. I do not recommend it for publication.
-
AC1: 'Reply on RC1', JIAYI JIN, 31 Jul 2021
The comment was uploaded in the form of a supplement: https://wes.copernicus.org/preprints/wes-2021-55/wes-2021-55-AC1-supplement.zip
-
RC2: 'Comment on wes-2021-55', Anonymous Referee #2, 04 Oct 2021
This paper used two approaches to estimate the AEP loss of a wind park in Norway due to icing.The first approach used a statistical investigation of the SCADA data, while the second approach compared the observed AEP with the results from a CFD study. The paper also tried to optimize the wind farm using the CFD tool, but this didn't seem to take into account the impact of icing, so it wasn't clear how that related to the main theme of the paper.
The main goal of the paper, to evaluate different approaches to estimating AEP loss is interesting, but the paper was quite challenging to follow and it isn't really clear to me what the main outcome of the study was.
Major comments
- The paper as a whole was challenging to read, there were many typos, and the grammer was not very good. The structure also was not laid out for the reader making it hard to follow what the recipe was to get to the end resuts.
- Several topics were only vaguely discussed and were difficult to understand. For example, there were several places where it was stated that the T19Method would be compared to the SCADA data, but the T19Method uses the SCADA data to get its results. I kept expecting there to be an ice detector or some other method to determine the icing periods or loss, but they weren't found.
- None of the methods are clearly described in such a way to allow the reader to be able to replicate the study. The T19IceLossMethod is probably the best described as it is the simplest method, but it is not totally clear how stops were detected, or if duration was used at all for determining icing periods. The Numerical model section (3.2) does not describe any of the input data, what was used to build the 3D terrain model, where did the roughness data come from, what was used for the inflow wind profile? In section 2, it is mentioned that a Weibull fit is carried out, but nothing about the method for fitting was described. While in all subsections in section 4 there were various methods applied that aren't described (4.1: how were the values of 57.1, 54.6, and 55.5% reached?; What is the % loss of total hour colume in table 6?; 4.2: Where can we see the vortex motion in figure 6 that you describe in this section?; 4.3 What is the T19IceLossMethod using in table and figure 7, so far all descriptions of that method produce results as % of ice loss, how do you then get an AEP?
- The data in the tables and figures was often hard to interpret and the reader was not guided to help understand it, rather than having discussion of the results, they were often just summarize in the text.
Minor comments:
Here I will mostly focus on the figures, tables, and references- Shouldn't the reference to the cold climate market either be Timo's presentation at winterwind 2017 (https://winterwind.se/wp-content/uploads/2015/08/9_1_24_Karlsson_IEA_Task_19_-_Cold_climate_wind_power_market_study_2015-2020_Pub_v1.pdf), or the Wind Power Monthly article (https://www.windpowermonthly.com/article/1403504/emerging-cold)?
- On line 38, you start of the paragraph as Gravdahl et al (Davis et al., 2016) it is unclear what the reference you were going for was, and the Gravdahl reference doesn't show up in the references.
- Table 1: it was unclear to me at first what was being show. I am now pretty confident that it contains the mean values for the 14 turbines, but it isn't clear what the reader should get from this table.
- Table 4: perhaps this information could be better portrayed through histograms of the different cell sizes, for example (https://link.springer.com/article/10.1007/s10546-020-00591-0/figures/3)
- Table 5: Is this the percentage of AEP loss? Something else?
- Table 6: If table 5 is correct, was there really 29.6% loss of AEP at turbine 13 with only 19.6 hours of down time, that seems quite extreme.
- Table 6: As mentioned above, there doesn't seem to be any description of that the % loss of total hour is, it seems like it might be the sum of the other columns by year, but should c be subtracted, and not all of the values sum correctly so I am a bit lost.
- Figure 5 is really hard to review, also it is known that there are a lot of points in certain parts of the power curve, consider instead using a hexbin plot like done in Figure 2 from (https://onlinelibrary.wiley.com/doi/10.1002/we.1878), for specific parts of the power curve you want to highlight. I was surprised to not see this paper referenced as it seems to be one of the closer papers to your work.
- Table 7: This seems a bit conterintuitive to me perhaps because in table 5 you were showing the percent loss, while here are showing the percent of the potential. It might be easier for the reader to keep using the same approach to describing the loss.
- Figure 7: What should be a better CFD result? I guess the closer the value is to the T19LossMethod? Since the SCADA data include icing effects that don't model in your CFD results.
- Figure 7: What does this figure add that you don't get from the table above?
Citation: https://doi.org/10.5194/wes-2021-55-RC2
Viewed
Total article views: 821 (including HTML, PDF, and XML)
Cumulative views and downloads
(calculated since 15 Jun 2021)
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 484 | 314 | 23 | 821 | 19 | 16 |
- HTML: 484
- PDF: 314
- XML: 23
- Total: 821
- BibTeX: 19
- EndNote: 16
Viewed (geographical distribution)
Total article views: 818 (including HTML, PDF, and XML)
Thereof 818 with geography defined
and 0 with unknown origin.
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
1
Cited
Latest update: 28 Apr 2024
Arctic Technology & Icing Research Group, Faculty of Engineering Science & Technology, UiT – The Arctic University of Norway, 8505 Narvik, Norway
Timo Karlsson
VTT Technical Research Centre of Finland, Espoo, Finland
Muhammad S. Virk
Arctic Technology & Icing Research Group, Faculty of Engineering Science & Technology, UiT – The Arctic University of Norway, 8505 Narvik, Norway
Short summary
In this manuscript, a numerical case study has been presented regarding ice detection and wind resource assessment in ice prone cold regions. Three years SCADA data from a wind park in Arctic region is used for this study. T19IceLossMethod based statistical analysis and Computational Fluid Dynamics (CFD) based numerical simulations are carried out for icing events classification and wind resource assessment, as well as estimation of resultant Annual Energy Production (AEP).
In this manuscript, a numerical case study has been presented regarding ice detection and wind...
Altmetrics