the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Prognostics-based adaptive control strategy for lifetime control of wind turbines
Edwin Kipchirchir
Manh Hung Do
Jackson Githu Njiri
Dirk Söffker
Abstract. Variability of wind profiles in both space and time is responsible for fatigue loading in wind turbine components. Advanced control methods for mitigating structural loading in these components have been proposed in previous works. These also incorporate other objectives like speed and power regulation for above-rated wind speed operation. In recent years, lifetime control and extension strategies have been proposed to guaranty power supply and operational reliability of wind turbines. These control strategies typically rely on a fatigue load evaluation criteria to determine the consumed lifetime of these components, subsequently varying the control set-point to guaranty a desired lifetime of the components. Most of these methods focus on controlling the lifetime of specific structural components of a wind turbine, typically the rotor blade or tower. Additionally, controllers are often designed to be valid about specific operating points, hence exhibit deteriorating performance in varying operating conditions. Therefore, they are not able to guaranty a desired lifetime in varying wind conditions. In this paper an adaptive lifetime control strategy is proposed for controlled ageing of rotor blades to guaranty a desired lifetime, while considering damage accumulation level in the tower. The method relies on an online structural health monitoring system to vary the lifetime controller gains based on a State of Health (SoH) measure by considering the desired lifetime at every time-step. For demonstration, a 1.5 MW National Renewable Energy Laboratory (NREL) reference wind turbine is used. The proposed adaptive lifetime controller regulates structural loading in the rotor blades to guaranty a predefined damage level at the desired lifetime without sacrificing on the speed regulation performance of the wind turbine. Additionally, significant reduction in the tower fatigue damage is observed.
Edwin Kipchirchir et al.
Status: final response (author comments only)
-
RC1: 'Comment on wes-2021-144', Anonymous Referee #1, 05 Jan 2022
This manuscript describes a combination of control to mitigate the loads with the objective of managing the lifetime of the wind turbine.
The authors have presented the methods and applied and demonstrated that for a mean wind speed of 18 m/s. However, 18 m/s is not the wind speed that would cause the most fatigue damage and the occurrence probability is relatively low. Therefore the authors should add/replace results from mean wind speeds that are more representative to demonstrate the capability of the controllers. It would be necessary to show the behavior of the wind turbine under rated wind speed. Here it can be seen how the controllers will perform in this transition region when the wind speed can be below rated and above rated. This can be used also to demonstrate and validate the switching behavior that the authors mentioned in the manuscript.
The switching behavior and implmentation in the controllers as mentioned in the manuscript should be described with more details. Another question that needs to be clarified is whether the constant switching of the controller will cause additional dynamics to the response of the wind turbine. Especially when the wind turbine is operating in the transition region.
The performance of the controller for the given turbine shown in Figure 8 of the manuscript seems to indicate that the rotor speed can deviate as much as 20% from the rated rotor speed (20 rpm) and the power can deviate more than 30% from the rated power (1500 kw). This is usually not possible as the overspeed protection will kick in as soon as the rotor speed is more than 110% of the rated rotor speed. The same would be applicable to the power, since the generator protection wlll kick in to protect the overheating of the generator. Therefore, the controller should be retuned to meet the standard performance requirements regarding overspeed and power deviation.
In Table 2 and in the text, the authors use the steady wind speed and prevailing wind speed in order to decide the switching of the controller that were tuned for different wind speeds. How are these wind speeds defined and how are they calculated in a continuous operation of the wind turbine, especially if one takes into account that the stationarity assumption of the wind does not really apply in reality.
The authors have considered the flapwise bending moment for the blade, while the edgewise bending moment play also an important role in the fatigue damage of the blade. One should consider the total bending moment of the blade for the estimation of the fatigue damage. The same should apply also to the tower fore-aft and side to side bending moment.
The wind field used for the validation of the method is not described sufficiently. It is not clear whether the stochastic wind field is coherent over the rotor plane and the quetstion remains whether one single realization of the stochastic wind field is representative enough to demonstrate the robustness of the controller.
Some additional remarks can be found on the attached pdf.
-
RC2: 'Comment on wes-2021-144', Anonymous Referee #2, 06 Apr 2022
The submitted paper proposes a new approach to include an optimized life time consumption calculation as an integrated part of a control strategy to mitigate the loads on rotor blades and tower.
After a very detailed description of standard fatigue calculation methods, the integrated control approach is presented, which is based on a Robust Disturbance Accommodating Controller (RDAC), published in previous papers by the authors. As a reference turbine, the NREL 1.5 MW model has been chosen, the simulation tool is FAST.
Linearization around several operating points above rated wind speed is proposed, for each of these points the controller is optimized, with switching mechanisms foreseen to allow a realistic operation under changing wind conditions.
The results of the controller implementation are presented for an average wind speed of 18 m/s and demonstrate that for both the blades and the tower the accumulated fatigue damage can be reduced simultaneously, claiming to have no negative effects on the power performance.
While the overall approach of this paper shows impressively the potential of improved controller schemes taking into account life time consumption, some details need to be clarified.
In the description of the NREL turbine models it is mentioned that the number of degrees of freedom is reduced, here the author should be more specific and explain their decision.
It is not clear what type of wind model is used and why the analysis is limited to just 18 m/s average wind speed. Showing the impact of more relevant lower wind speeds around rated and demonstrating the switching mechanism would be interesting.
It is pretty obvious that directly related load components as flap-wise for the blade and fore-aft for the tower correlate in their behavior. Also taking into account the edgewise loads and the related tower movements would complete the picture.
The very high dynamics of the torque/speed signal need to be explained.
The baseline control strategy for the comparisons needs to be described in more detail – is it RDAC with or without IPC?
To compensate for some more additional results, the introduction can be shortened by referring to standard literature instead of explaining in detail the basics of fatigue calculation.
Some spelling errors should be eliminated, e.g. guarantee instead of guaranty etc.
Citation: https://doi.org/10.5194/wes-2021-144-RC2 -
RC3: 'Comment on wes-2021-144', Anonymous Referee #3, 19 Apr 2022
Discussion of:
Prognostics-based adaptive control strategy for lifetime control of wind turbines
10 April 2022
The paper presents a suggested adaptive control strategy that could be applied to limit the fatigue damage accumulation in selected wind turbine components, for the wind speed range where the turbine control is based on pitch regulation. The paper shows how the suggested controller strategy successfully limits the loads in a few scenarios, however it fails to show the overall significance of the new strategy with respect to the entire operating envelope of the wind turbine, and does not show any quantitative assessment of the impacts of applying the suggested strategy. A new version of this paper would need to focus significantly more on the validation and performance evaluation of the suggested strategy. A few clarifying comments are below:
General comments
- I don’t think the paper title is correct. There are no prognostics discussed in the paper whatsoever, it is rather load mitigation. Hence I would instead call it “Adaptive control strategy for load-based lifetime consumption control of wind turbines”
- It is hard to judge the practical significance of this method. It works only for wind speeds above 12m/s, which in reality only occurs about 25% of the time on a typical site.
- I suspect that if this approach is also applied at lower wind speeds, the power output may be reduced. These and any other limitations need to be clearly outlined.
- There is no quantitative assessment of the performance of the suggested procedure. How much exactly are the loads reduced, what is the increase in the pitch actuator duty cycles, is the behaviour robust and consistent over different realizations? This needs to be shown both for individual wind speeds, but also the total effect over the turbine lifetime needs to be estimated.
- The English needs some checks - there are some spelling issues to correct like “guaranty” instead of “guarantee” but also others.
Citation: https://doi.org/10.5194/wes-2021-144-RC3 - AC1: 'Comment on wes-2021-144', Edwin Kipchirchir, 16 May 2022
Edwin Kipchirchir et al.
Edwin Kipchirchir et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 440 | 183 | 13 | 636 | 7 | 8 |
- HTML: 440
- PDF: 183
- XML: 13
- Total: 636
- BibTeX: 7
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1