Quantification of parameter uncertainty in wind farm wake modeling

Abstract

Please refer to published version for the most recent bibliographic citation information. If a published version is known of, the repository item page linked to above, will contain details on accessing it.

A novel dynamic wind farm wake model based on deep learning

“…These models are formulated analytically and are very fast to evaluate, which makes them suitable for wind farm layout optimization. The further development of analytical models is still an active area, such as taking into account the turbulence effect [12,13], modelling yaw effects [14,15,16], modelling background flow effects [17], considering the expansion of physical wake boundary [18], and incorporating uncertainty based on high-fidelity data [19]. However, as these models are static, they are mainly used in turbine layout design for optimizing static quantities, such as mean power generation.…”

Optimizing wind farms layouts for maximum energy production using probabilistic inference: Benchmarking reveals superior computational efficiency and scalability

“…This confidence levels on the energy production can be complemented by quantifying uncertainty of both input parameters [78,79,80] and wake models [81,82], which are likely to have a more significant impact in the resulting optimal layout. A comprehensive robust optimization approach that considers and quantifies all these uncertainties, although beyond of the scope of the present study, would be a worthy goal for future work in this area.…”

Wind farm yaw control set-point optimization under model parameter uncertainty

“…The probability distributions of the wake model parameters can be estimated using approximate Bayesian inference methods, such as Markov chain Monte Carlo 22,25 . However, sampling based methods are computationally intensive, requiring O(10 5 ) forward model evaluations to perform the Bayesian inverse problem analysis 25 .…”

“…Recent work has optimized the wake model parameters using only wind farm power data and analytic gradients 5 , a novel calibration procedure 19 , genetic algorithms 20 , and Kalman filtering 21 and demonstrated that assimilating operational wind farm data into the wake model improves its predictive capability. Zhang & Zhao (2020) 22 used sampling to approximate the Bayesian posterior distributions of wake model parameters given LES data as the ground truth. Using the wake model parameter posteriors, a stochastic wake model based on uncertain model parameters was proposed which improved predictions compared to wake modeling with deterministic model parameters.…”