A Meandering-Capturing Wake Model Coupled to Rotor-Based Flow-Sensing for Operational Wind Farm Flow Estimation by Maxime LEJEUNE

A numerical investigation

Wind Farm Flow Control (WFFC) refers to the coordinated control of the turbines inside wind farms with the goal of influencing the flow in such a way that it improves the global performances of the wind power plant. Gaining improved flow awareness is consequently paramount to enhancing the overall effectiveness and robustness of WFFC. Different avenues toward better flow awareness can be be pursued: (i) closed-loop control, (ii) improved flow-sensing techniques, (iii) more accurate, yet efficient, flow modeling. This thesis primarily focuses on the latter two approaches which are nonetheless prerequisites to efficient closed-loop control.

More specifically, it leverages Large Eddy Simulations (LES) of wind farms to support the development, tuning and testing of an operational dynamic flow modeling framework aimed at capturing the dynamic signature of the flow at a low computational cost while relying only on information gathered at the wind turbine location. The ensuing framework, named OnWaRDS, brings together flow sensing and Lagrangian flow modeling. The features of the inflow are first inferred from their blade load response: a Kalman filter coupled to a Blade Element Momentum theory solver is used to determine the rotor-normal flow velocity while an Artificial Neural Net trained on high-fidelity numerical data estimates of the rotor-transverse wind velocity component. The information recovered is in turn propagated in a physics-informed fashion across the domain by the Lagrangian flow model which contributes to incrementally reconstructing an estimated snapshot of the full wind farm flow field.

The ensuing framework is first presented and then deployed within a numerical wind farm where its performances are assessed. Comparison against the LES baseline reveals that the proposed model achieves good estimates of the flow state under various operating and atmospheric conditions. The model distinctly provides additional insight into the wake physics when compared to the traditional steady state approach: the characteristic signature of wake meandering along with the influence of large-scale fluctuations of the ambient flow are captured based solely on the information measured by the wind turbines. This analysis further confirms the computational affordability and high modularity of the proposed framework in line with the requirements of WFFC.

Jury members :

• Prof. Philippe Chatelain (UCLouvain, Belgium), supervisor
• Prof. Aude Simar (UCLouvain, Belgium), chairperson
• Prof. Grégoire Winckelmans (UCLouvain, Belgium)
• Prof. Julien Hendrickx (UCLouvain, Belgium)
• Dr. Thufe Göçmen (DTU, Denmark)
• Prof. Jeroen van Beeck (VKI, Belgium)

Join on your computer, mobile app or room device

Click here to join the meeting

Meeting ID: 311 005 081 788
Passcode: 4wf9CP