Affordable Access

Access to the full text

Evaluation of Anomaly Detection of an Autoencoder Based on Maintenace Information and Scada-Data

Authors
  • lutz, marc-alexander
  • vogt, stephan
  • berkhout, volker
  • faulstich, stefan
  • dienst, steffen
  • steinmetz, urs
  • gück, christian
  • ortega, andres
Publication Date
Feb 29, 2020
Identifiers
DOI: 10.3390/en13051063
OAI: oai:mdpi.com:/1996-1073/13/5/1063/
Source
MDPI
Keywords
Language
English
License
Green
External links

Abstract

The usage of machine learning techniques is widely spread and has also been implemented in the wind industry in the last years. Many of these techniques have shown great success but need to constantly prove the expectation of functionality. This paper describes a new method to monitor the health of a wind turbine using an undercomplete autoencoder. To evaluate the health monitoring quality of the autoencoder, the number of anomalies before an event has happened are to be considered. The results show that around 35% of all historical events that have resulted into a failure show many anomalies. Furthermore, the wind turbine subsystems which are subject to good detectability are the rotor system and the control system. If only one third of the service duties can be planned in advance, and thereby the scheduling time can be reduced, huge cost saving potentials can be seen.

Report this publication

Statistics

Seen <100 times