An advanced farm flow estimator for the real-time evaluation of the potential wind power of a down-regulated wind farm

Abstract

This work aims at verifying the predictions of OnWaRDS, an open-source wake modelling framework that captures the main features of the wake dynamics, including its meandering, in ancillary services scenarios. OnWaRDS brings together Lagrangian flow modeling and flow sensing and runs in parallel with a wind farm environment (here Large-Eddy simulations coupled to an Actuator Disk model (LES-AD)) in order to use the available rotor states to predict the flow field. The performances of OnWaRDS are first assessed when it runs synchronously with the LES-AD of a down-regulated wind farm and tends to mimic the LES-AD behavior. This synchronous mode implies a continuous feeding of the wake model with LES-AD rotor data. Then, OnWaRDS is used in a predictive mode, in order to predict an alternate reality for the wind farm. In this study, OnWaRDS aims at evaluating, in real-time, what the potential power production would be when the LES-AD is down-regulated to provide operating reserve capacity to the electricity network. Switching to a predictive mode implies that certain measurements at the wind turbine level can no longer be used, because the flow and the rotor behavior change between LES-AD and OnWaRDS. The second part of this study thus aims at verifying the predictions of OnWaRDS, and highlighting the impact of switching from a synchronous to a predictive mode in OnWaRDS.