Implementation and Analyses of Yaw Based Coordinated Control of Wind Farms

Abstract

This paper presents, with a live field experiment, the potential of increasing wind farm power generation by optimally yawing upstream wind turbine for reducing wake effects as a part of the SmartEOLE project. Two 2MW turbines from the Le Sole de Moulin Vieux (SMV) wind farm are used for this purpose. The upstream turbine (SMV6) is operated with a yaw offset ( α ) in a range of − 12 ° to 8° for analysing the impact on the downstream turbine (SMV5). Simulations are performed with intelligent control strategies for estimating optimum α settings. Simulations show that optimal α can increase net production of the two turbines by more than 5%. The impact of α on SMV6 is quantified using the data obtained during the experiment. A comparison of the data obtained during the experiment is carried out with data obtained during normal operations in similar wind conditions. This comparison show that an optimum or near-optimum α increases net production by more than 5% in wake affected wind conditions, which is in confirmation with the simulated results.

Wind Farm Yaw Optimization via Random Search Algorithm

“…In conjunction with these wake models and wind farm data, control algorithms can be used to optimize yaw angles. Several studies have developed optimization-based control approaches to adjust to the uncertainties, using approaches such as Bayesian optimization [41], genetic algorithm [42], game theory [43,44], random search [45], sequential optimization [32], gradient descent [28,46], greedy control [47], particle swarm optimization [48], dynamic programming [49,50]. Most recently, Howland et al [37] developed an control scheme based on an empirical fitted analytical wake model [26,40].…”

Dynamic wake modulation induced by utility-scale wind turbine operation

“…However, the wake is not a passive tracer; rather, the turbine actively modulates the incoming flow to influence wake behavior. Controlling wakes using yaw error alone (Ahmad et al 2019) and a combination of both axial induction and yaw error (Munters & Meyers 2018) has also been proposed as a method for optimizing overall wind farm performance. However, to effectively implement these strategies, we need to first understand the impact on wake behavior of the constantly-changing flow and turbine operational conditions present in the field.…”

Exploring the complexities associated with full-scale wind plant wake mitigation control experiments

“…The nacelle anemometer (subject to the NTF) was used to denote the turbine inflow wind speed and define the blade pitch angle offsets. However, nacelle-based measurements are inherently distorted due to their location behind the rotating rotor (Allik et al, 2014). Furthermore, the NTF is a proverbial black box; it is unknown what turbine signals besides the nacelle anemometer are used to produce the turbine inflow wind speed, and it is also unknown how well it approximates rotor-sweep-relative variations in wind speed.…”

“…Experimental validation of wind plant control at full scale has frequently relied upon the analysis of power and controls data from individual turbine pairs in a wind plant to quantify the benefit of various wind plant control techniques (e.g., Fleming et al, 2017aFleming et al, , 2019Ahmad et al, 2019;Van der Hoek et al, 2019;Howland et al, 2019). However, few studies have used advanced measurement technologies (such as lidar or radar) to document differences in wake structure due to the turbine control changes implemented as part of wind plant control (e.g., Trujillo et al, 2016;Fleming et al, 2017b).…”