Active wake control: An approach to optimize the lifetime operation of wind farms

Abstract

The wind turbines within a wind farm impact each other's power production and loads through their wakes. Wake control strategies, aiming to reduce wake effects, receive increasing interest by both the research community and the industry. A number of recent simulation studies with high fidelity wake models indicate that wake mitigation control is a very promising concept for increasing the power production of a wind farm and/or reducing the fatigue loading on wind turbines' components. The purpose of this paper is to study the benefits of wake mitigation control in terms of lifetime power production and fatigue loading on several existing full-scale commercial wind farms with different scale, layouts, and turbine sizes. For modeling the wake interactions, Energy Research Centre of the Netherlands' FarmFlow software is used: a 3D parabolized Navier-Stokes code, including a kturbulence model. In addition, an optimization approach is proposed that maximizes the lifetime power production, thereby incorporating the fatigue loads into the optimization criterion in terms of a lifetime extension factor. KEYWORDS active wake control, fatigue loads reduction, power production maximization, wake mitigation, wind farm control INTRODUCTIONWind turbines are being densely clustered in wind farms and effect each other's power production and loading through their wakes. Because of the reduced wind velocity and increased turbulence, a turbine operating in the wake of another turbine will produce less power and experience increased fatigue loading. Still, the common practice is to let each wind turbine maximize its individual power capture, thereby disregarding its effect on other turbines. This "greedy" approach is not optimal with respect to the power production of the whole wind farm. At the beginning of this century, researchers started the development of a cooperative approach to operate wind farms 1-3 that aim at maximizing the power production of the whole farm, while at the same time, trying to reduce the fatigue loading on the wind turbines. These methods for mitigation of wake effects are called here active wake control (AWC).There are 2 major classes of AWC methods. The first concept aims at reducing the wake deficit downstream by reducing the axial induction factor of upstream wind turbines, known as axial induction control (also referred to as heat and flux or pitch-based AWC). This is achieved by operating the turbines at the windward side at an increased blade pitch angle. 4,5 In practice, increased pitch angle could indirectly be achieved by reducing the power production of the turbine (derating). While the power production of the derated machines decreases, the wind velocity in their wakes increases, allowing to produce more power at downstream wind turbines, and, possibly, raises the overall farm production. Initial experiments in a wind tunnel 5,6 and in the field 7 have shown that the power production of a wind farm may potentially increase under pitch-based AWC. However, more-recent high-fidelity simulations 8 a...

Massive simplification of the wind farm layout optimization problem

“…10. In addition to testing wind farms with different turbine spacing, we modeled and optimized several different wind farm boundaries in this study: the boundary of the Princess Amalia wind farm, a real farm in the North Sea (Van Dam et al, 2012;Gebraad and Van Wingerden, 2015;Kanev et al, 2018), a circle, and a square to demonstrate the sharp angles that can occur in wind farm boundaries. These boundaries are shown in Fig.…”

An active power control approach for wake-induced load alleviation in a fully developed wind farm boundary layer

“…The increased turbulence intensity inside the wakes, wake meandering, and partial wake overlapping of the rotors of downwind turbines cause waked turbines to be prone to higher structural fatigue loading. Several studies have utilized optimization techniques to find optimal set-points in order to simultaneously maximize the total power production and prolong the lifespan of the wind farm through minimizing the structural fatigue loading (van Dijk et al, 2017;Kanev et al, 2018). Nonetheless, from a control engineering perspective, these have been either open-loop or quasisteady-state optimization approaches, based on analytical static wake models and data-driven load models, which do not fully hold for dynamical wake interactions.…”

A tutorial on the synthesis and validation of a closed-loop wind farm controller using a steady-state surrogate model

“…Among others, [8], [9] demonstrated the concept in highfidelity simulation. Furthermore, [10], [11] 1 There is research towards model-free methods for wind farm optimization (e.g., [5]- [7]), but the time delays involved in wake propagation pose a real challenge to such methods. This is not further explored here, and the interested reader is referred to Boersma et al [4].…”