the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 30 Aug 2023
OC6 Project Phase IV: Validation of Numerical Models for Novel Floating Offshore Wind Support Structures
Abstract. This paper provides a summary of the work done within Phase IV of the Offshore Code Comparison Collaboration, Continued, with Correlation and unCertainty (OC6) project, under International Energy Agency Wind Technology Collaboration Programme Task 30. This phase focused on validating the loading on and motion of a novel floating offshore wind system. Numerical models of a 3.6-MW horizontal-axis wind turbine atop the TetraSpar floating support structure were compared using measurement data from a 1:43 Froude-scale test performed in the University of Maine’s Alfond Wind-Wave (W2) Ocean Engineering Laboratory. Participants in the project ran a series of simulations, including system equilibrium, surge offsets, free-decays, wind-only conditions, wave-only conditions, and a combination of wind and wave conditions. Validation of the models was performed by comparing the aerodynamic loading, floating support structure motion, tower base loading, mooring line tensions, and keel line tensions. The results show a good estimation of the aerodynamic loading and a reasonable estimation of the platform motion and tower base fore-aft bending moment. However, there is a significant dispersion in the dynamic loading for the upwind mooring line. Very good agreement was observed between most of the numerical models and the experiment for the keel line tensions.
- Preprint
(2457 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on wes-2023-103', Carlos Silva de Souza, 18 Sep 2023
The paper provides a comparison between numerical simulations and model test results for a floating wind turbine, adopting the "TetraSpar" substucture. The simulations were carried out by 17 participants, adopting different software and assumptions for the hydrodynamic loads, aerodynamic loads, structural model, and mooring system formulation. The comparisons are made in terms of relevant quantities in the analysis of floating wind turbines, including horizontal platform motions, mooring lines tension, and tower base bending moment. In addition, the tensions on the lines connecting the platform upper structure and the keel are also assessed.
The topic has great relevance, since previous comparisons have shown significant discrepancies between different simulation software. However, the authors mentioned important sources of inaccuracy in the comparison, the most serious of them being the influence of the model's "umbilical" in the floater mean position, and on wind drag loads during idling conditions. In addition, the initial position of the floater changed significantly throughout the test campaign, which directly affects the mean mooring lines loads.
It can be noted that the platform wave-frequency response varies significantly in amplitude for the different participants. The authors are strongly encouraged to better discuss this point, since the response for platforms with a more conventional design is normally well predicted at the wave-frequency range. In particular, the discussion should consider the model for the hydrodynamic loads (Morison's equation, potential flow theory, hybrid) adopted by the different participants.
A PDF file containing details/suggestions for improvement will be sent to the authors directly.
-
AC2: 'Reply on RC1', Roger Bergua Archeli, 30 Dec 2023
Dear Carlos Silva de Souza,
Thank you for taking the time to review the manuscript and provide useful feedback.
We have rewritten part of the manuscript to address the comments made during this review process.
Attached you can find a PDF with replied to the comments.
The revised version of the manuscript where these comments are included will be uploaded soon.
-
AC3: 'Reply on RC1', Roger Bergua Archeli, 30 Dec 2023
Dear Carlos Silva de Souza,
Thank you for taking the time to review the manuscript and provide useful feedback.
We have rewritten part of the manuscript to address the comments made during this review process.
Attached you can find a PDF with replied to the comments.
The revised version of the manuscript where these comments are included will be uploaded soon.
-
AC2: 'Reply on RC1', Roger Bergua Archeli, 30 Dec 2023
-
RC2: 'Comment on wes-2023-103', Anonymous Referee #2, 27 Sep 2023
In this paper, the results of the OC6 code-to-experiment comparison for the TetraSpar was presented, with focus on the responses of floater motions, tower bottom bending moment, keel line and mooring line tensions. The paper is very comprehensive. Some of the notable discrepancies could be further investigated.
In Table 3, the authors can also check how the buoyancy force was modelled when using the Morison model. Is the buoyancy force calculated for the instantaneous position of the floater or at the mean position?
For example, for the wave-frequency motion responses, it seems that DUT2-SIMA give much higher results. Usually the wave-frequency responses can be well captured. Probably there are some modelling errors. It could be related to the double count of the first order wave loads, using the Morison model, in which the potential theory might also be used. It could also be related to the buoyancy calculation of the slender elements in the Morison model.
Some codes give very high structural responses of the tower or the keel/mooring line tensions, at high frequencies. It might be related to the fact that the structural damping of these elements is not properly considered.
Regarding the cases with wind, it is very difficult to regenerate the wind field and the wind thrust time series in the lab test as requested for model testing. However, were the wind loads at the tower top measured so that the external wind thrust force time series can be derived by removing the turbine inertial loads? Then, they can be applied to the numerical models for comparison. Or, was the wind field measured and used in the numerical codes? Or, the numerical codes generate their own wind field and wind loads on turbines?
Citation: https://doi.org/10.5194/wes-2023-103-RC2 -
AC1: 'Reply on RC2', Roger Bergua Archeli, 30 Dec 2023
Dear Referee #2,
Thank you for taking the time to review the manuscript and provide constructive feedback. This helped us to have a more complete and comprehensive manuscript.
In Section 3 (Participants and Modeling Approach), we have included in the new manuscript a description about how the buoyancy force was modeled by the different participants:
Some participants used a linear hydrostatic stiffness while others used a nonlinear hydrostatic stiffness. Participants using the PF theory (NU) and some participants using a hybrid approach (BVMO), included a linear hydrostatic stiffness computed at the undisplaced platform position. Some participants using the hybrid approach (EDF, TUHH) or the strip theory (DUT1, DUT2), computed the nonlinear hydrostatics at the instantaneous platform position up to the mean sea level. Most participants using the strip theory (CSSC, DNV, DTU, GDG, IFE, MAR1, NREL, PRI, SHELL, UPC, W&UG) and CFD (MAR2) computed the nonlinear hydrostatics at the instantaneous platform position up to the wetted surface.
We agree that there might be a problem with the numerical model used by DUT2-SIMA. According to the description provided by the participant, they use a Morison equation approach. It is not a hybrid model (i.e., there is no potential flow body). Similar issues can be observed for DUT2-SIMA for the fairlead 2 tension and the keel line tensions.
Regarding the high structural responses for the tower, in the original manuscript there were some patricipants (e.g., GDG) with large responses in the first tower bending mode. Many other participants obtained similar tower responses along the project. We understand that this is a limitation of Morison-only models. Our understanding is that in the relatively high frequency of the first tower bending mode (around 0.4 Hz), the Morison approach is able to excite the floating system with a loading that is not realistic. Most participants were able to obtain the expected loads after low-pass filtering the irregular wave spectrum (recommended practive DNV-RP-C205) or applying the MacCamy and Fuchs diffraction correction of the inertia coefficient. For reference, the definition document that the participants used to build up the numerical models includes information about the structural damping for the tower (obtained from hammer impact tests).
For the high response of the mooring line tensions (e.g., DUT2-SIMA), this could be a problem with the structural damping of the line or lack of hydrodynamic damping (i.e., drag coefficient). We have added one paragraph in the conclusions to summarize the issues that have been commented along the manuscript:
It is important to note that it was the first time that project participants built numerical models for this novel floating support structure design. Due to the multi-physics nature of the system, including aerodynamics, hydrodynamics and structural dynamics, participants had to use a coupled aero-hydro-elastic approach. The different participants defined a modeling approach according to the capabilities of the code(s) used. Setting up these numerical models was quite challenging and prone to user error. This is evidenced by the fact that some participants used the same code and a similar modeling approach but obtained different system responses.
Regarding the cases with wind, in the experimental campaign there was no load sensor at the tower top. Only one triaxial accelerometer was installed at tower top. The wind in the wave basin was measured during the calibration process by means of an ultrasonic anemometer (near the hub height location) and one hotwire anemometer (located upwind, slightly at one side and lower than the hub height). During the testing, the ultrasonic anemometer was removed, but the hotwire was still in place to ensure that the wind generated by the fans was the expected one. For the numerical models, participants considered spatially uniform unsteady winds based on the measured hub-height wind speed in the X-direction. This means that they took the recorded wind at the hub height and assumed that the same wind was applied over the whole rotor plane. For reference, the turbulence intensity during the testing was quite low.Citation: https://doi.org/10.5194/wes-2023-103-AC1
-
AC1: 'Reply on RC2', Roger Bergua Archeli, 30 Dec 2023
Status: closed
-
RC1: 'Comment on wes-2023-103', Carlos Silva de Souza, 18 Sep 2023
The paper provides a comparison between numerical simulations and model test results for a floating wind turbine, adopting the "TetraSpar" substucture. The simulations were carried out by 17 participants, adopting different software and assumptions for the hydrodynamic loads, aerodynamic loads, structural model, and mooring system formulation. The comparisons are made in terms of relevant quantities in the analysis of floating wind turbines, including horizontal platform motions, mooring lines tension, and tower base bending moment. In addition, the tensions on the lines connecting the platform upper structure and the keel are also assessed.
The topic has great relevance, since previous comparisons have shown significant discrepancies between different simulation software. However, the authors mentioned important sources of inaccuracy in the comparison, the most serious of them being the influence of the model's "umbilical" in the floater mean position, and on wind drag loads during idling conditions. In addition, the initial position of the floater changed significantly throughout the test campaign, which directly affects the mean mooring lines loads.
It can be noted that the platform wave-frequency response varies significantly in amplitude for the different participants. The authors are strongly encouraged to better discuss this point, since the response for platforms with a more conventional design is normally well predicted at the wave-frequency range. In particular, the discussion should consider the model for the hydrodynamic loads (Morison's equation, potential flow theory, hybrid) adopted by the different participants.
A PDF file containing details/suggestions for improvement will be sent to the authors directly.
-
AC2: 'Reply on RC1', Roger Bergua Archeli, 30 Dec 2023
Dear Carlos Silva de Souza,
Thank you for taking the time to review the manuscript and provide useful feedback.
We have rewritten part of the manuscript to address the comments made during this review process.
Attached you can find a PDF with replied to the comments.
The revised version of the manuscript where these comments are included will be uploaded soon.
-
AC3: 'Reply on RC1', Roger Bergua Archeli, 30 Dec 2023
Dear Carlos Silva de Souza,
Thank you for taking the time to review the manuscript and provide useful feedback.
We have rewritten part of the manuscript to address the comments made during this review process.
Attached you can find a PDF with replied to the comments.
The revised version of the manuscript where these comments are included will be uploaded soon.
-
AC2: 'Reply on RC1', Roger Bergua Archeli, 30 Dec 2023
-
RC2: 'Comment on wes-2023-103', Anonymous Referee #2, 27 Sep 2023
In this paper, the results of the OC6 code-to-experiment comparison for the TetraSpar was presented, with focus on the responses of floater motions, tower bottom bending moment, keel line and mooring line tensions. The paper is very comprehensive. Some of the notable discrepancies could be further investigated.
In Table 3, the authors can also check how the buoyancy force was modelled when using the Morison model. Is the buoyancy force calculated for the instantaneous position of the floater or at the mean position?
For example, for the wave-frequency motion responses, it seems that DUT2-SIMA give much higher results. Usually the wave-frequency responses can be well captured. Probably there are some modelling errors. It could be related to the double count of the first order wave loads, using the Morison model, in which the potential theory might also be used. It could also be related to the buoyancy calculation of the slender elements in the Morison model.
Some codes give very high structural responses of the tower or the keel/mooring line tensions, at high frequencies. It might be related to the fact that the structural damping of these elements is not properly considered.
Regarding the cases with wind, it is very difficult to regenerate the wind field and the wind thrust time series in the lab test as requested for model testing. However, were the wind loads at the tower top measured so that the external wind thrust force time series can be derived by removing the turbine inertial loads? Then, they can be applied to the numerical models for comparison. Or, was the wind field measured and used in the numerical codes? Or, the numerical codes generate their own wind field and wind loads on turbines?
Citation: https://doi.org/10.5194/wes-2023-103-RC2 -
AC1: 'Reply on RC2', Roger Bergua Archeli, 30 Dec 2023
Dear Referee #2,
Thank you for taking the time to review the manuscript and provide constructive feedback. This helped us to have a more complete and comprehensive manuscript.
In Section 3 (Participants and Modeling Approach), we have included in the new manuscript a description about how the buoyancy force was modeled by the different participants:
Some participants used a linear hydrostatic stiffness while others used a nonlinear hydrostatic stiffness. Participants using the PF theory (NU) and some participants using a hybrid approach (BVMO), included a linear hydrostatic stiffness computed at the undisplaced platform position. Some participants using the hybrid approach (EDF, TUHH) or the strip theory (DUT1, DUT2), computed the nonlinear hydrostatics at the instantaneous platform position up to the mean sea level. Most participants using the strip theory (CSSC, DNV, DTU, GDG, IFE, MAR1, NREL, PRI, SHELL, UPC, W&UG) and CFD (MAR2) computed the nonlinear hydrostatics at the instantaneous platform position up to the wetted surface.
We agree that there might be a problem with the numerical model used by DUT2-SIMA. According to the description provided by the participant, they use a Morison equation approach. It is not a hybrid model (i.e., there is no potential flow body). Similar issues can be observed for DUT2-SIMA for the fairlead 2 tension and the keel line tensions.
Regarding the high structural responses for the tower, in the original manuscript there were some patricipants (e.g., GDG) with large responses in the first tower bending mode. Many other participants obtained similar tower responses along the project. We understand that this is a limitation of Morison-only models. Our understanding is that in the relatively high frequency of the first tower bending mode (around 0.4 Hz), the Morison approach is able to excite the floating system with a loading that is not realistic. Most participants were able to obtain the expected loads after low-pass filtering the irregular wave spectrum (recommended practive DNV-RP-C205) or applying the MacCamy and Fuchs diffraction correction of the inertia coefficient. For reference, the definition document that the participants used to build up the numerical models includes information about the structural damping for the tower (obtained from hammer impact tests).
For the high response of the mooring line tensions (e.g., DUT2-SIMA), this could be a problem with the structural damping of the line or lack of hydrodynamic damping (i.e., drag coefficient). We have added one paragraph in the conclusions to summarize the issues that have been commented along the manuscript:
It is important to note that it was the first time that project participants built numerical models for this novel floating support structure design. Due to the multi-physics nature of the system, including aerodynamics, hydrodynamics and structural dynamics, participants had to use a coupled aero-hydro-elastic approach. The different participants defined a modeling approach according to the capabilities of the code(s) used. Setting up these numerical models was quite challenging and prone to user error. This is evidenced by the fact that some participants used the same code and a similar modeling approach but obtained different system responses.
Regarding the cases with wind, in the experimental campaign there was no load sensor at the tower top. Only one triaxial accelerometer was installed at tower top. The wind in the wave basin was measured during the calibration process by means of an ultrasonic anemometer (near the hub height location) and one hotwire anemometer (located upwind, slightly at one side and lower than the hub height). During the testing, the ultrasonic anemometer was removed, but the hotwire was still in place to ensure that the wind generated by the fans was the expected one. For the numerical models, participants considered spatially uniform unsteady winds based on the measured hub-height wind speed in the X-direction. This means that they took the recorded wind at the hub height and assumed that the same wind was applied over the whole rotor plane. For reference, the turbulence intensity during the testing was quite low.Citation: https://doi.org/10.5194/wes-2023-103-AC1
-
AC1: 'Reply on RC2', Roger Bergua Archeli, 30 Dec 2023
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 777 | 354 | 23 | 1,154 | 14 | 13 |
- HTML: 777
- PDF: 354
- XML: 23
- Total: 1,154
- BibTeX: 14
- EndNote: 13
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1