Active Wind Farm Cluster Wake Mixing (Clusterwake)

WHY

In the North Sea, wind turbines are arranged in large offshore clusters, so-called wind farms. Currently, 30 GW of wind energy has been installed, which should be expanded to 120 GW by 2030 and 300 GW by 2050. Clustering wind turbines in wind farms has numerous benefits, such as lower infrastructure and maintenance costs. But there are also negative effects, like the wake effect. Because a turbine extracts energy from the wind, the wind downstream of the turbine becomes less energetic.

With the growing number of wind farms, it is becoming clear that the wake effect is present not only at the level from the turbine to the turbine but also at the level from the wind farm to the wind farm. At the wind farm level, energy extraction by each individual turbine cumulates, leaving a clear wake in the downstream area. As a result, the affected downstream wind farms will experience significant power losses.

Given the high density of planned wind farms in the North Sea, the impact of wind farm wake losses on European wind energy production could become substantial. Despite this realisation, the wind turbine cluster wake effect is still not fundamentally well understood. Currently, no strategy mitigates the cluster wake-up effect. In contrast, the wake of individual wind turbines has been extensively studied. The current state of the art includes several strategies to mitigate the effects on downstream turbines, including accelerating wake recovery, by dynamically exciting the pitch angles of the blades for example, known as Active Wake Mixing (AWM).

TU Delft recently presented an initial concept around Active Cluster Wake Mixing (ACWM). Active Cluster Wake Mixing generates dynamic force patterns across the wind farm by adjusting the thrust of the individual turbines. These patterns cause instability in the wind farm wake. Given the scale of a wind farm, the excitation frequency is significantly lower than the frequencies used to break down the wake of a single turbine, so only minor effects on the fatigue loads of the individual turbines are expected. Initial calculations show a power gain in the wind farm itself of 2%, while the reduction of cluster wake effects on neighbouring wind farms is estimated to be up to 30%.

To realise these gains, we propose varying the thrust of each wind turbine to generate unstable, non-uniform spatial and temporal thrust distribution patterns over the entire wind farm. The resulting gradual variation in thrust force generates additional vorticity, leading to increased wake mixing and cluster wake instability.

WHAT

In this project, we want to make progress on three fundamental topics:

1. The theory of what causes cluster wake mixing,
2. The patterns that most effectively drive wake mixing and
3. Efficient ways to test, evaluate and validate interactions between farms.

We will try to explain why we observe a structurally earlier wake breakdown due to increased mixing. We will also use a variety of computational fluid dynamics (CFD)-based methodologies to gain valuable insights into the governing mechanisms. In this project, we will develop a versatile toolbox to generate excitation protocols. To do this, we need to classify known patterns and their effects. Where shown to be feasible, we will use control techniques to optimise the operational strategy.

We will test, evaluate and validate actuation patterns through simulations and experiments. We will conduct simulations by further developing and validating the so-called Actuator Farm models. Recent results enable the simulation of turbines at coarser grid resolutions with lower computational costs. Experimental work on a laboratory scale will allow rapid testing of different actuation patterns. To this end, we are building on existing work in wind farm modelling to develop porous disks that can be actively monitored.

Finally, we use the interaction between wind farms as a test case. Based on these outcomes, we will identify suitable wind farms for this technology and explore the potential benefits for those wind farms. We will use the results of previous work for comparison and to demonstrate the potential of the novel technology. We will work with the data of these wind farms to make the case study as realistic as possible so that a field test campaign can be set up in parallel.

EXPECTED RESULTS

This project presents the first step of the development roadmap of this promising innovation in wind farm control. At the end of this project, we will better understand the potential of Active Cluster Wake Mixing (ACWM). Current computer models will have been adapted and improved to estimate cluster wake effects properly. We will use realistic data from existing wind parks in the North Sea to validate the models.

Through this project, we will learn what good strategies are for upwind wind farms to generate additional energy yields in downwind wind farms.