Μεταπτυχιακές Εργασίες

Μόνιμο URI για αυτήν τη συλλογήhttps://pyxida.aueb.gr/handle/123456789/7

Περιήγηση

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Α Β Γ Δ Ε Ζ Η Θ Ι Κ Λ Μ Ν Ξ Ο Π Ρ Σ Τ Υ Φ Χ Ψ Ω

Τώρα δείχνει 1 - 1 από 1

Gaussian processes for modelling wind turbine power curves
(12/03/2021) Ladopoulou, Domniki; Λαδοπούλου, Δομνίκη; Athens University of Economics and Business, Department of Informatics; Karlis, Dimitrios; Zois, Georgios; Dellaportas, Petros; Vassalos, Vasilios
Modelling wind turbine power curves is a hugely important problem in management and monitoring wind power on wind farms. Such models are commonly utilised (i) to report to the authorities the energy that the wind farm will deliver to the national provider, (ii) to provide a tool to monitor the good functionality/operation of each wind turbine by comparing the expected and observed output power, (iii) to evaluate the wind farm development in a location in which only weather characteristics have been measured and (iv) to impute missing data of a wind turbine that has temporarily stopped without deleting the data of the other turbines. We provide a novel approach based on multi-task Gaussian processes that provides probabilistic answers to the above questions and we demonstrate that it outperforms single-task Gaussian process modelling. Our final model specification predicts the joint density of power curves for given wind speeds, wind directions, temperature and air pressure.