Field testing of multi-variable individual pitch control on a utility-scale wind turbine
Introduction
Active control techniques are an indispensable part of modern wind turbines. Historically, control strategies enabled the improvement of wind turbine technology to the point where wind turbines are aerodynamically and structurally operated optimally at each wind speed. Specifically, state-of-the-art components of modern utility scale wind turbines include torque controlled generators to achieve optimal aerodynamic operation, yaw control to direct the turbine into the wind, collective pitch control to ensure safe operation in high wind scenarios, power electronic controllers to decouple the turbine operation from the power grid as well as different load reduction controllers Thus, control techniques form the basis for improving wind turbine power generation.
Clearly, the future will bring larger scale wind turbines both on- and offshore with an increased power production to meet society’s demands on clean energy, see, e.g., Refs. [[1], [2], [3]]. Control techniques can again help to overcome the problems coming along with such large-scale structures. One of those problems is the increased flexibility due to the high dimensions of the rotor. The problem is intensified by the demand for saving material and building the blades with less material. In addition, modern and future turbine rotors cover huge areas making them prone to turbulence as well as that the rotation results in a kind of self-excitation of the structure (e.g., when passing the tower shadow) results in unwanted oscillations and loads of the structure. Modern control techniques can help to alleviate the resulting oscillations and loads on the turbine and thereby increase efficiency of power generation and reduce operational costs. One possibility is the well-known individual pitch control (IPC) approach to allow for the damping of unwanted out-of-plane oscillations and thus, loads, in the blades. The available IPC literature is substantial and provides different approaches to solve the IPC problem. Most of them are somehow related to the ideas presented in the early IPC work from the 2000’s [[4], [5], [6]], while unfortunately only this early work has been actually field tested on real turbines and made available to the research community. Besides IPC, active flaps are becoming an interesting option for wind turbine load reduction which have been successfully field tested on a utility-scale turbine in Ref. [7].
Generally, a wind turbine’s dynamical properties depend on the rotary hub position, making it a periodic dynamic system. To overcome the issue that well developed linear control techniques cannot be applied to periodic systems, the multi blade coordinate (MBC) transformation described in Refs. [8,9], has been introduced in wind turbine control. It projects rotating quantities onto non-rotating coordinates. This provides signals and mathematical models in the non-rotating frame as detailed in Refs. [10,11]. The classical IPC strategy described in Refs. [6,12] commonly processes the MBC transformed blade loads by two single-input-single-output (SISO) integral or proportional-integral feedback controllers in order to suppress low frequency components of the loads. The enabling of sophisticated, model-based control approaches lead to developments of multi-variable control designs based on a linear quadratic regulator in Ref. [4] and based on -norm optimal control in Refs. [13,14]. These controllers aimed at the once-per-revolution (1P) blade loads and were shown to yield similar results as the classical IPC strategy.
More recent approaches also try to reduce loads at higher frequencies as they also reduce the lifetime of turbine. A direct extension of the two SISO loop strategy is a model free approach discussed in Refs. [[15], [16], [17], [18]] additionally using complex, higher-order transformations of the sensor data. However, in this case, the analysis of a single, linear closed loop model to verify robustness and performance is not straightforward anymore. Further, it requires several design steps instead of a single one. Model-based, multi-variable design approaches to target higher frequency loads on the blades have been contributed by the authors in Ref. [19] for a 2.5 MW turbine using control as well as linear parameter-varying control in Ref. [20]. Designing IPC load reduction algorithms for higher frequency inevitably calls for multi-variable-control approaches as the cross-coupling between the transformed channels cannot be neglected. Commonly multi-loop SISO control approaches simply fail to achieve a stable closed loop system due to cross-coupling in the channels as demonstrated in Ref. [19]. Specially for bigger turbines where the load reduction at higher frequencies will be required due to the increased flexibility, the multiple SISO loop strategy may be insufficient.
Thus, to tackle load reduction problems on the increasing wind turbine structures in the future, multi-variable controllers are a possible solution as they are based on well-developed control design methodology. However, such control techniques still lack industrial acceptance and are not used frequently for wind turbine load reductions. The reason for this is, in the views of the authors.
- (a)
that current state-of the art turbines do not fully demand the load reductions at higher frequencies offered by such controllers but also
- (b)
that these methods are still considered to be an academic methodology.
The latter represents a gap between academia and industry. This gap arises from a general lack of actual field test validations of novel control contributions by academia. Motivated by the fact that (a) does not hold any more for the next generation of very large wind turbines this article aims to overcome (b) by presenting the results of a field campaign of a multi-variable control design for the 2.5 MW Clipper Liberty C96 research turbine. This available turbine for which the control system is modified is operated by the University of Minnesota and presents a state-of-the art utility scale wind turbine.
A description of the Clipper Liberty C96 research turbine is provided including a discussion of the available non-linear open loop model, the baseline controller as well as a comprehensive analysis of the derived linear design models (Section 2). The -based IPC control algorithm itself, designed to tackle the 1P loads on the blades like classical IPC, is also presented in Section 2. The article provides insights in the step-wise approach from the model-based design to the final IPC integration, which allowed a safe test of the advanced control technique on the utility scale turbine (Section 3). Finally, the results of a field test campaign comparing the -based IPC with no-IPC as well as classical IPC are presented (Section 4).
Section snippets
Individual blade pitch controller design for the Clipper Liberty
The wind turbine considered in this article is the utility-scale three-bladed Clipper Liberty 2.5 MW research turbine of the Eolos Wind Energy Research Field Station located at the UMore Park in Rosemount, MN, and shown in Fig. 1. It has a hub-height of 80m and a rotor diameter of 96m. The operating range of wind turbines is commonly classified into three regions based on rotor speed: standstill (region 1), variable speed operation (region 2), and constant speed operation (region 3). The
IPC integration
The complete architecture of the Clipper Liberty’s control system is depicted in Fig. 5. The architecture includes the baseline controller along with the proposed individual blade-pitch controller for load reduction. The baseline controller controls generator torque for an aerodynamic optimal operation in region 2 via a so-called -law and maintains rated rotor speed via collective blade pitch in region 3 operation. The collective blade pitch command in region 3 is calculated through
Field tests results
The field test campaign for which the data herein is presented was performed between March and September 2019. The goal was to generate data allowing a comparison of the two load controllers against each other and against the no-IPC scenario, i.e., using collective blade pitch control only. Therefore, in each test run the three controller setups have been run consecutively to enable a meaningful comparison in similar wind conditions. Depending on the availability of the turbine as well as the
Conclusions
The field test results of a novel wind turbine control algorithm at the utility scale 2.5 MW Clipper Liberty C96 wind turbine to reduce structural loads have been presented. The highlights of the performed research can be summarized as follows:
- •
With the presented multi-variable control design approach the coupling effects between the yaw and pitch axis in the fixed frame, which are commonly neglected in the classical IPC strategy, can be explicitly considered to improve load reduction
CRediT authorship contribution statement
Daniel Ossmann: Controller development, simulation and field test data analysis, Writing - original draft. Peter Seiler: Formal analysis, Resources, Project administration, Writing - review & editing. Christopher Milliren: Wind turbine hardware upgrade, field test logistics and execution, Writing - review & editing. Alan Danker: Wind turbine hardware upgrade, controller code implementation, Writing - review & editing.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was performed in the framework of the Xcel Energy Renewable Energy Fund: Contract Number RD4-13. The project title is Virtual Wind Simulator with Advanced Control & Aeroelastic Model for Improving the Operation of Wind Farms.
References (30)
- et al.
Downwind coning concept rotor for a 25 MW offshore wind turbine
Renew. Energy
(2020) - et al.
State-space formulae for all stabilizing controllers that satisfy an H∞-norm bound and relations to risk sensitivity
Syst. Contr. Lett.
(1988) - et al.
Simple rainflow counting algorithms
Int. J. Fatig.
(1982) - et al.
Aerodynamic design of the 13.2 MW SUMR-13i wind turbine rotor
- M. Kaminski, C. Noyes, E. Loth, R. Damiani, S. Hughes, C. Bay, M. Chetan, D. T. Griffith, K. Johnson, D. Martin,...
Individual blade pitch control for load reduction
Wind Energy
(2003)Further load reductions with individual pitch control
Wind Energy
(2005)- et al.
Controller Field Tests on the NREL CART2 Turbine, Tech. Rep. NREL/TP-5000-49085
(2010) - et al.
Field test of an active flap system on a full-scale wind turbine
Wind Energy Science
(2021) - et al.
Theory of Self-Excited Mechanical Oscillations of Helicopter Rotors with Hinged Blades, Tech. Rep. NACA-TR-1351
(1958)
Helicopter Theory
Stability analysis of a variable-speed wind turbine
Multi-blade coordinate transformation and its application to wind turbine analysis
Developments in closed loop controller design for wind turbines
Cited by (23)
Toward control co-design of utility-scale wind turbines: Collective vs. individual blade pitch control
2023, Energy ReportsCitation Excerpt :However, to improve the competitiveness of wind energy with respect to other carbon-derived sources of energy, further enhancements are required to decrease the LCOE of wind farms. Efforts to reduce the LCOE have been directed toward increasing the wind turbines energy production and reducing loads through advanced control systems (Bitar and Seiler, 2013; Rotea, 2014; Gebraad and van Wingerden, 2015; Johnson and Fritsch, 2012; Mirzaei et al., 2015; Ciri et al., 2017b,a, 2018; López et al., 2020; Bossanyi, 2003; Geyler and Caselitz, 2008; Vali et al., 2016; Ossmann et al., 2021). Recent attempts to increase power production have been aimed at the overall efficiency of wind farms.
Peak shaving strategy for load reduction of wind turbines based on model predictive control
2023, Energy ReportsCitation Excerpt :Load reduction is another task of the operation controller. The control objective can be vibration velocity (Mourad and Ieee, 2019), force (Vlaho and Bottasso Carlo, 2017), or moment (Daniel et al., 2021). Additional electromagnetic torque can eliminate the torsional vibration of the drive train (Liang et al., 2022), and the pitch angle can be appended to increase the damping of the fore-aft tower motion (Haoming et al., 2020).
Periodic LQG Wind Turbine Control with Adaptive Load Reduction
2023, IFAC-PapersOnLineActuator fault tolerant offshore wind turbine load mitigation control
2023, Renewable EnergyA robust gain scheduling method for a PI collective pitch controller of multi-MW onshore wind turbines
2022, Renewable EnergyCitation Excerpt :Furthermore, a feed-forward loop was introduced in Ref. [17], where a real 3-bladed turbine built in association with NREL (National Renewable Energy Laboratory) and the University of Stuttgart was tested using a LIDAR (Light Detection and Ranging) system to send wind data ahead of time to avoid the aforementioned errors. Recently, the 2.5 MW Clipper Liberty research turbine was field tested based on a MIMO (Multi-input Multi-output) IPC control scheme for structural load reduction [18]. This work helped better investigate how advanced controllers operate in real life compared to simulations.