the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Feb 2025
System identification of offshore wind turbines for model updating and validation using field measurements
Abstract. This study presents an applied system identification approach for developing, updating, and validating simulation models of wind turbines using field measurements. This is demonstrated by developing a model of a bottom-fixed offshore turbine in the Belgian North Sea. An initial model is obtained based on available design information and the scaling of reference models. Afterwards, the model is calibrated by leveraging drivetrain vibration data and blade strain measurements. This is accomplished using Operational Modal Analysis techniques, which enable the identification of the turbine’s modal parameters, particularly its eigenfrequencies. By identifying the eigenmodes, the model can be updated to match the modal behaviour of the deployed turbine. Comparison with SCADA data further validates the model’s operational performance. Additionally, the obtained model is used to calculate blade pitch-bearing lifetime estimations in different wind conditions, to give insight in the detrimental effect of wind properties on the pitch-bearing lifetime, which can be used as decision support in operation and maintenance strategies for wind turbines.
Competing interests: Amir R. Nejad is member of the editorial board of the Wind Energy Science journal. The authors declare that they have no further conflicts of interest.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.- Preprint
(2237 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 10 Apr 2025)
-
RC1: 'Comment on wes-2024-173', Anonymous Referee #1, 12 Mar 2025
reply
Dear authors, thank your for your submission titled "System identification of offshore wind turbines for model updating and validation using field measurements". Your work is relevant and it is worth being published in the Wind Energy Science journal. The article is fairly well written and the scientific approaches are appropriate. I do have a list of comments. Some are minor, whereas others are likely to require some more substantial changes.
Major comments:
1) The article is the combination of two fairly distinct investigations. In particular, the tuning of the model described in the first part is likely quite irrelevant to the second part where the lifetime of the bearings is estimated. In the first part of the article you tune the elastic properties of blades and tower. These do not seem to be the key drivers of loads in the pitch bearings. Is it a good idea to combine the two stories together? Please discuss.
2) Related to point #1, the power curve comparison presented in section 2.4.1 is not really a validation. If you showed the initial basic model in Figure 8, I believe it would overlap with the red markers generated by the tuned model. The power curve is only mildly influenced by the natural frequencies of the system (except if major instabilities are present, but that's probably not your case). So I don't think this section should be part of section 2.4 Validation. The only real validation is presented in Section 2.4.2.
3) The paper only has 4 sections. I think readability would improve significantly if you could split the narrative into more sections. One idea could be to do: Section 2 Model Generation Section 3 System Identification Section 4 Model Tuning Section 5 Model Validation. If you do so, remember to update the text at line 78.
4) The whole scaling process is pretty crude, but, more importantly, it is not well documented. At line 107, what does it mean that "tower bending stiffness is scaled based on rated thrust"? Same for line 113: what does it mean "scaled accordingly"? Section 2.1.3 is also nebulous. I understand that data is confidential, but why do you use such a complex tool like QBLade to do a Viterna extension of the polars? What does it mean that "material properties are linearly scaled"? Or "thickness ratio is equal to 1"? It seems that you simply scaled blade mass by the cube of the length. Am I missing anything? Overall, I would recommend a substantial rewriting of these sections, adding some rigor to the description of the scaling process.
5) Table 1: at 0 rpm you should not have rotor whirling modes. Why do you have 3 distinct natural frequencies for flap and 3 for edge? Also, I don't think you discuss how you've estimated these numbers. The comment also applies to line 426: I don't think you've obtained the first 13 eigenfrequencies. I think it's 7.
Minor comments:
- The introduction is too verbose. You submitted your manuscript to a wind energy journal, there is no need to talk about global deployment goals. Also because these get very quickly outdated: the USA have just pivoted away from offshore wind (your citation at line 16 is outdated). Please rethink the text between lines 13 and 40 and make it more specific to your manuscript. Please focus on your contributions that are of interest to your readers.
- Line 68: there is a major difference between damage equivalent loads and lifetime estimates, which are usually based on stress metrics. Aero-servo-hydro-elastic models estimate DELs, but for lifetime more is needed. Please highlight this critical difference.
- Line 85: "around" 8MW? I imagine rated power is not proprietary...
- Line 86: This is maybe not of critical importance, but it's surprising to see an expensive commercial tool such as Simpack being used solely for visualization and modal characterization... Although in a clunkier fashion, but NREL tools do both things.
- Line 104: the LEANWIND model could not be validated, when validation means compared to real-world data, since the LEANWIND model was purely theoretical. Note that the citation Desmond et al. 2013 is misleading, because it points to a DNV deliverable that I cannot find online (is it even publicly available?). I would recommend rephrasing this paragraph.
- Line 116: what does "fixity" mean?
- Line 117: define DOF. Also, I think you can better explain that 0 degrees of freedom mean rigid clamping.
- Line 170: it's not entirely clear why you used 4 turbines and not just one. What is the reasoning and value of using 4? For example, the results do not clear characterize turbine to turbine variability.
- Line 175: I find confusing that you interchange out-of-plane with edgewise and in-plane with flapwise. I understand that the blades are parked, but I would stick to the words "edgewise" and "flapwise".
- Line 191: typo "Utilizing"
- Line 207: what kind of sensors are you using? did you install them, or did they come installed from the manufacturer? These are important details for replicability. For example, can your approach be replicated on any turbine, or does it require the installation of specific instrumentation?
- Line 209: typo "Hz"
- Line 214: typo "measurements"
- Figure 6: I don't see where you've defined how you've normalized Frequency
- Line 229: I think I understand what a "yaw-inducing" mode is, but I am less sure about the "pitch-inducing" mode? Maybe better to link it back to Figure 5?
- Line 236: typo "strain"
- Line 328: "as well"? Where else are these parameters specified?
- Line 334: "The movement is a slow oscillation, which makes the periodicity that the classical calculation approach builds on disappear." What does this mean?
- Line 405: "Due to turbulence, the wind speed varies over time, when it surpasses the rated speed, the pitch angle is constant at 0 deg, thus reducing the number of pitch movements." What does this mean? Same for "where the thrust is still high but the rated wind speed is seldom surpassed."
Thank you for letting me review your work and best regards
Citation: https://doi.org/10.5194/wes-2024-173-RC1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 116 | 29 | 4 | 149 | 6 | 5 |
- HTML: 116
- PDF: 29
- XML: 4
- Total: 149
- BibTeX: 6
- EndNote: 5
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1