EphèseLe catalogue de la bibliothèque de l'ECAM

Titre :	Anomaly detection on wind turbines components, before they reach a dangerous level, with deep learning methods applied on SCADA dataset
Type de document :	Travail de fin d'études
Auteurs :	Hubert Mugabo Pitie, Auteur ; Dimitrios Tourtounis, ; François Defrance,
Editeur :	ECAM
Année de publication :	2019
Note générale :	AKKA TECHNOLOGIES
Langues :	Anglais (eng)
Index. décimale :	TFE - Informatique (ECAM)
Résumé :	The transition from fossil energies to renewable ones will require the use of different energy sources. Wind Turbine (WT) is one of the solutions as it uses the wind (a renewable source) as the energy source. The WT costs come mainly from the operating cost, the cost to maintain the WT. Today, the maintenance is done periodically or after a problem is detected. To reduce this cost, artificial intelligence methods (Neural Networks) can be used with the Supervisory Control and Data Acquisition (data) system to monitor the condition of Wind Turbine components. The monitoring is done by predicting anomalous behavior. The goal of this thesis is to implement a predictive maintenance, via Neural Networks, to detect anomalous behaviors and prevent them or reduce their impact.

Anomaly detection on wind turbines components, before they reach a dangerous level, with deep learning methods applied on SCADA dataset [Travail de fin d'études] / Hubert Mugabo Pitie, Auteur ; Dimitrios Tourtounis, ; François Defrance,  . - ECAM, 2019.
AKKA TECHNOLOGIES
Langues : Anglais (eng)

Index. décimale :

TFE - Informatique (ECAM)

Résumé :

The transition from fossil energies to renewable ones will require the use of different energy sources. Wind Turbine (WT) is one of the solutions as it uses the wind (a renewable source) as the energy source. The WT costs come mainly from the operating cost, the cost to maintain the WT. Today, the maintenance is done periodically or after a problem is detected. To reduce this cost, artificial intelligence methods (Neural Networks) can be used with the Supervisory Control and Data Acquisition (data) system to monitor the condition of Wind Turbine components. The monitoring is done by predicting anomalous behavior. The goal of this thesis is to implement a predictive maintenance, via Neural Networks, to detect anomalous behaviors and prevent them or reduce their impact.