Skip to Main content Skip to Navigation
Theses

Adaptive Multi-Agent Systems for Wind Power Forecasting

Tanguy Esteoule 1
1 IRIT-SMAC - Systèmes Multi-Agents Coopératifs
IRIT - Institut de recherche en informatique de Toulouse
Abstract : The use of renewable energies, particularly wind power, is one of the solutions commonly used to limit the worsening of ongoing climate change. The variability and intermittency of these energy sources are the main constraints to be managed to ensure the integration of renewable energies into the electricity grid. This problem can be partly solved by improving production forecasts in the short and medium term. The theory of AMAS (Adaptive Multi-Agent Systems) proposes to solve complex problems by self-organization for which no algorithmic solution is known. The local and cooperative behavior of the agents allows the system to adapt to a dynamic environment for maintaining the system in an adequate operating state. In this thesis, this approach is applied to the forecasting of wind farm production. More specifically, we are studying the integration of finer scale data (wind farms for a region or wind turbines for a farm) into the forecast model. We therefore propose a method that takes into account local data in the global forecast and more precisely the interdependencies between wind turbine and wind farm productions. The study led to the design of two adaptive multi-agent systems: AMAWind-Turbine forecasting the production of a wind farm using wind turbine data, and AMAWind-Farm forecasting the production of a region using wind farm data. These systems have been tested in real conditions on five wind farms currently in operation. The experiments carried out validated the proper functioning of the systems and showed a decrease in forecasting error, the main factor in the field of application.
Document type :
Theses
Complete list of metadata

Cited literature [241 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02930244
Contributor : Abes Star :  Contact
Submitted on : Friday, September 4, 2020 - 11:51:06 AM
Last modification on : Thursday, June 10, 2021 - 3:07:24 AM
Long-term archiving on: : Wednesday, December 2, 2020 - 8:27:21 PM

File

2019TOU30246b.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02930244, version 1

Citation

Tanguy Esteoule. Adaptive Multi-Agent Systems for Wind Power Forecasting. Artificial Intelligence [cs.AI]. Université Paul Sabatier - Toulouse III, 2019. English. ⟨NNT : 2019TOU30246⟩. ⟨tel-02930244⟩

Share

Metrics

Record views

129

Files downloads

422