Abstract
This paper describes the use of a combination of anomaly detection and data-trending techniques encapsulated in a multi-agent framework for the development of a fault detection system for wind turbines. Its purpose is to provide early error or degradation detection and diagnosis for the internal mechanical components of the turbine with the aim of minimising overall maintenance costs for wind farm owners. The software is to be distributed and run partly on an embedded microprocessor mounted physically on the turbine and on a PC offsite. The software will corroborate events detected from the data sources on both platforms and provide information regarding incipient faults to the user through a convenient and easy to use interface.
| Original language | English |
|---|---|
| Title of host publication | 2007 IEEE Lausanne Powertech Proceedings |
| Place of Publication | New York |
| Publisher | IEEE |
| Pages | 22-27 |
| Number of pages | 6 |
| ISBN (Print) | 9781424421893 |
| DOIs | |
| Publication status | Published - 2007 |
| Event | IEEE Lausanne Powertech - Lausanne, Switzerland Duration: 1 Jul 2007 → 5 Jul 2007 |
Conference
| Conference | IEEE Lausanne Powertech |
|---|---|
| Country/Territory | Switzerland |
| City | Lausanne |
| Period | 1/07/07 → 5/07/07 |
Keywords
- power generation economics , Power system economics
- wind turbines
- wind farms
- sensor systems
- artificial intelligence
- degradation
- fault detection
- fault diagnosis
- intelligent sensors
- multi-agent
- fault detection system
- wind turbine
- defect recognition
- diagnosis