Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Deployment of Multi-Agent-Systems for optimising O&M of wind turbine

Presentation held at 8th PhD Seminar on Wind Energy in Europe September 12-14, 2012, ETH Zurich, Switzerland
 
: Rafik, K.; Faulstich, S.; Pfaffel, S.; Kühn, P.

:
Volltext urn:nbn:de:0011-n-3938484 (1.1 MByte PDF)
MD5 Fingerprint: 9622dd0c20d1cb4e981d23b0dfc1743c
Erstellt am: 8.6.2016


2012, 10 S.
PhD Seminar on Wind Energy in Europe <8, 2012, Zurich>
Englisch
Vortrag, Elektronische Publikation
Fraunhofer IWES ()

Abstract
Maintenance management for wind turbines (WT) aims on the one hand at reducing the overall maintenance cost and on the other hand at improving the availability.Although modern onshore WT attain high technical availability of up to 98 %, the evaluation of maintenance work in previous projects shows, that high WT availability requires additional maintenance work and costs. There is a considerable scope for optimizing reliability and maintenance procedures. A possibility therefore is to systematically make use of available knowledge and past experience. At this point, information coming from databases, statistical methods as well as sound statements are essential. The consideration of several conditions e.g. weather conditions, power forecasts, stock keeping etc. are essential for optimal maintenance decisions. However, due to this enormous amount of information sophisticated tools are needed. The contribution will present the possible application of high-performance computing methodologies, which may help wind farm operators (WFO) examining optimal maintenance strategies. The so called Multi-Agent-System (MAS) which is a new discipline in the world of Artificial Intelligence (AI) and the Data Mining (DM), which is a high-performance computing methodology used to observe and deduce hidden knowledge and logical dependencies of a great amount of data using several appropriate algorithms, should be investigated and a methodology for the use of AI in WT maintenance is proposed.

: http://publica.fraunhofer.de/dokumente/N-393848.html