Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Data driven condition monitoring of wind power plants using cluster analysis

: Li, Peng; Eickmeyer, Jens; Niggemann, Oliver

Preprint urn:nbn:de:0011-n-3643024 (986 KByte PDF)
MD5 Fingerprint: 3588b8223b77a4515e06c6022383d253
© IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Created on: 4.11.2015

Institute of Electrical and Electronics Engineers -IEEE-:
International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery, CyberC 2015. Proceedings : 17. Sep.-19. Sep. 2015, Xi'an, China
Piscataway, NJ: IEEE, 2015
ISBN: 978-1-4673-9199-3
International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery (CyberC) <7, 2015, Xi'an>
Conference Paper, Electronic Publication
Fraunhofer IOSB ()

Along with the rapid growth of the wind energy sector, reducing the wind energy costs caused by unplanned downtimes and maintenances has aroused great concern of researchers. Condition monitoring system (CMS) is widely used for detecting anomalies of wind power plants (WPPs) so as to reduce the downtimes and optimize the maintenance plan. However, current solutions to condition monitoring of WPPs focus mostly on detecting a particular anomaly on a single component or a subsystem. Optimizing the maintenance plan of whole wind power plant requires a solution to system level condition monitoring of WPPs. This paper gives an approach for system level condition monitoring of WPPs using data driven method, that provides an overall picture of the system statuses. Firstly, cluster analysis is utilized to automatically learn the normal behavior model of WPPs from the observations. Two clustering algorithms are explored to choose a suitable one for modeling the WPPs. The presented anomaly detection algorithm uses the learned model as reference to detect the system anomalies. The effectiveness of this approach is evaluated with real world data.