|
Title:
|
|
Conmow Final Report
|
|
|
|
Author(s):
|
Wiggelinkhuizen, E.J.; Rademakers, L.W.M.M.; Verbruggen, T.W.; Watson, S.J.; Xiang, J.; Giebel, G.; Norton, E.; Tipluica, M.C.; Christensen, A.J.; Becker, E.
|
|
|
|
Published by:
|
Publication date:
|
|
ECN
Wind Energy
|
22-6-2007
|
|
|
|
ECN report number:
|
Document type:
|
|
ECN-E--07-044
|
ECN publication
|
|
|
|
Number of pages:
|
Full text:
|
|
72
|
Download PDF
|
Abstract:
For offshore turbines the demand for high reliability and low operation and maintenance (O&M) costs is higher than for onshore turbines. Recent studies show that the O&M costs of offshore wind farms are too high, about 25 to 30% of the energy generation costs, and a considerable part is caused by unexpected failures leading to corrective maintenance.
In the CONMOW project it has been investigated whether online condition monitoring (CM-)techniques have added value for optimizing O&M strategies of large offshore wind farms and how these techniques can be improved.
Therefore first an inventory has been made of available and suitable state-of-the-art CMtechniques followed by a failure mode and effect analysis to select meaningful (applications for)condition monitoring systems and to identify useful developments.
Then a measurement campaign was carried out on a GE 1.5S turbine and on five Nordex N80/2.5 MW turbines to assess and improve the performance of CM-techniques. Results from five different drive train vibration monitoring systems have been analysed in combination with high-frequency data from “traditional” measurement systems and from the turbine PLC. By analysing the high-frequency electric power signal using wavelet transformations a shaft misalignment could be detected. This was confirmed by vibration measurements, which provided more detailed information on the origin and the effects of these vibrations.
Further large amounts of SCADA statistical data and inspection reports have been analysed. Several methods to process and present these data, such as de-trending and filtering, showed to reduce scatter and resulted in a simple and orderly presentation to the operator. This makes it easier to observe trends and abnormalities, which can improve early failure detection. However, due to the limited measurement time during which no failures occurred in the turbines, no proof was found that the methods indeed are useful to determine failures at an early stage.
The presently available drive train monitoring systems, supplied by Gram&Juhl and Pr??ftechnik performed well and reliable. It was demonstrated that the systems are able to detect component errors at an early stage as well as off-design conditions, such as shaft misalignment. Up till now,
the response on abnormalities in the data is either more frequent inspections or an immediate shut down to avoid consequence damage, as yet insufficient knowledge is available to make prognoses how the failures will develop in order to change from calendar based maintenance to condition based maintenance. Such knowledge can only be obtained from a larger population of identical wind turbines and longer measurement periods during which faults occur.
In the project several improvements have been implemented in the vibration monitoring systems, including more accurate vibration measurements on the low-frequency stages, improved configuration and processing techniques and automated reporting via Internet.
All types of measurement systems and CM-systems produce large data amounts which are difficult to interpret by wind farm operators, although reporting has improved. Therefore one is dependent on dedicated experts to derive meaningful recommendations for O&M optimization.
Instead of the original ambition to implement the lessons learnt into a wind farm SCADA system and test these over a longer period of time, additional simulations have been performed. Simulations with GH Bladed showed that in some cases fault conditions and off-design conditions can be detected from measurements. A cost sensitivity study showed a clear effect of the quality of the CM-systems, e.g. percentage of false alarms or non-detected failures, on the potential O&M cost reduction. Both the ECN cost model as the Risø cost model clearly showed that significant cost benefits are possible when suitable CM-techniques are applied
Back to List