Skip to main content

Closed Loop Wind Farm Control

Periodic Reporting for period 2 - CL-Windcon (Closed Loop Wind Farm Control)

Reporting period: 2018-05-01 to 2019-10-31

Current practice in wind farm operations is to control each turbine individually with regard to energy capture and loads based on locally available measurements. Commercial wind farms are typically operated then at an individual turbine level. Wake interactions and their effect on power production are usually taken into account during wind farm planning, design and commissioning, but not taken into account while operating the wind farms, except for specific O&M problems. At a wind farm level there is a lot of unleveraged potential with regard to overall operational optimisation by taking into account all available data across the farm as well as farm-level interactions among turbines.

The project has provided the wind energy sector, with a new control technology and a new market for business, since CL-Windcon has proposed introducing a totally new marketable technology which will improve the overall wind farm performance.

CL-Windcon has developed open-source software tools, with control algorithms validated in simulations, wind tunnel and field test. The project has unleashed a reduction of LCoE relevant enough to the industry’s expectations, and improvements of turbine and farm-level reliability and availability. CL-Windcon has included a comprehensive analysis of economic and environmental impact of the technical improvements resulting from the project and a review of standards for future wind turbine and wind farm design.

The project has sought integrated wind farm optimisation, by using individual turbine as actuators at wind farm level, acting wind turbines as a subordinated system to improve the performance of the wind farm. The actuation of the wind turbine can be done in three different ways: derating upwind turbines, yawing upwind turbines or activating IPC of downwind turbines

To this end, CL-Windcon has developed advanced control algorithms for axial induction and wake redirection that optimise the operation of the wind farm for a balance between annual energy production, life, and O&M cost, aimed at minimising lifetime LCoE. The project has applied techniques including loads-optimised power curtailment, event-triggered IPC for loads reduction under partial wake conditions, as well as fault-tolerant and fast wake recovery techniques.
The project has defined reference wind-farm layouts and simulation cases, which have been shared with the IEA Task 37, specifically dealing with reference definitions in wind energy.

A set of multi-fidelity wind farm models has been developed. A classification of selected tools’ capabilities has been performed, providing open guidelines for proper application. Additionally, a complete high-fidelity reference wind farm simulation framework has been developed in SOWFA. The framework and detailed documentation is available open access to the wind energy community and has been widely disseminated among target professionals through different means such as a training workshop at the Wind Energy Science Conference (WESC, 2019), the Topical Expert Meeting on Wind Farm Control promoted in clustering with the IEA-Wind Task 11, NREL and FarmConners European project.
Important steps have been taken to provide turbine controllers with the algorithms and software they need to, in turn, provide farm controllers with means to derate or change the yaw of upwind turbines or modify the pitch of downwind turbines. Steps have also been taken to provide turbine controllers with turbine health information and increase their room for manoeuvre according to said information by allowing turbines to operate in the face of certain failures. Moreover, wake and wind state estimators, as well as a method that enables the online update of wind farm models, have been developed.

The project has deal not only with theoretical analysis, but also with experimental tests to increase TRL of the technology. Wind tunnel test experiments were executed focused on the characterisation of wind turbine wakes under different operation conditions, as well as on the investigation of different control strategies. The results obtained up to this point confirm that the proposed control and load mitigation strategies are effective, and the learnings will be incorporated in the testing of integrated wind-turbine/wind-farm controllers.

In addition, field test campaign has also been performed for wake model validations, induction control and wake redirection control. Important results has been obtained, confirming the most important conclusions coming from simulations and wind tunnel tests, although research community needs to put more effort on this kind of validations, which are complex and expensive.

All developed control solutions in CL-Windcon have been published on Github, specifically; the optimisation tools for the FLORIS model are published and are already being used by multiple partners inside and outside of the project. Additionally, the state estimation algorithm for the dynamic surrogate model has already been used in different experiments outside of the CL-Windcon project, with positive responses.
CL-Windcon has covered the ambition of moving from greedy behaviour of wind turbines, to system behaviour, by moving from theoretical static control to open-loop wind farm control and also towards dynamic closed-loop control. The wind farm controller needs to adapt its control actions in real time to optimise the farm operation to the time-varying wake interaction.

In addition to dynamic farm control, the project has developed supporting control strategies on the turbine level:
- the event-driven IPC activated when partial wake operation is detected, significantly reducing the pitch actuator duty, and therefore the fatigue loading on the pitch bearing and transmission, while improving loads on the wind turbine only when it is needed.
- the loads-optimised power set-point tracking, which enables to follow external reference for the active power but this is achieved in such a way that the fatigue loading is minimised.

The economic ambition is to achieve a significant improvement of the operational efficiency of wind farms in terms of increased power production, reduced O&M costs, increased availability and lifetime extension. The expected outcome of this project shows a LCOE reduction of 0.63% under the most conservative models and scenarios, and for only using one of the proposed solutions. This result is aligned with market end users expectation.

In terms of environment, the project will complete an LCA assessment to evaluate the overall value chain of the new control strategy. The CO2 emissions removal on the base of the project potential impact is esteemed in more than 4.5 Mio tons between years 2020 and 2025.

Regarding socio-economic impact, the project will generate green business opportunities and jobs, fostering the European competitiveness and increasing the public acceptance; job creation over 2800 direct employments from 2020 to 2025.

CL-Windcon is also improving design norms affecting the loads and power performance of individual wind turbines by integrating new designs of the wind farm and the individual wind turbines, by means of the creation of a set of pre-normative and legal issues (standards) for wind turbine and /wind farms design affected by wind farm control algorithms.
Wind tunnel tests