KAStrion was a project entitled “Current and vibration analysis for preventive and predictive condition-based maintenance in wind farms”. It was fund by the KIC InnoEnergy from 2012 to 2014. The aim of this paper is to sum up and highlight the main results of the project. KAStrion goals were to maximize the production time of wind turbine farms by delivering a complete solution build upon a stand-alone analysis system which delivers a continuous on-site pre-diagnostic of the machine based on a multi-modal spectral monitoring technology. This embedded system located in the nacelle is connected to a tailored diagnostic center which delivers a periodic reporting on technical state of each machine of the farm. The strong innovation of KAStrion was to develop firstly a data-driven signal processing, referred to as AStrion, to automatically analyze, detect, classify all the spectral structures (harmonics and sidebands) of vibration signals, and secondly an original approach, referred to as SMESA, to process polyphase electrical signals. Contrary to existing systems, the coupling with the system kinematics is done after the analysis. KAStrion system has been tested on a specific test bench designed as a wind turbine at a smaller scale with load units on the main bearing, the planetary gear box and the output bearing in order to generate defects within an endurance test program. When compared with standard condition monitoring features, KAStrion shows its ability to characterize the start and the stage of the fault without the need of a historical data base. KAStrion system is also continuously tested on 2 two wind turbines in Arfons windfarm in France
[1]
Corinne Mailhes,et al.
Monitoring based on time-frequency tracking of estimated harmonic series and modulation sidebands
,
2014
.
[2]
Corinne Mailhes,et al.
AStrion data validation of non-stationary wind turbine signals Topic: Condition monitoring (CM) methods and technologies
,
2015
.
[3]
Waldo Vinton Lyon,et al.
Transients analysis of alternating-current machinery : an application of the method of symmetrical components
,
1954
.
[4]
Marcin Firla,et al.
Automatic and Full-Band Demodulation for Fault Detection—Validation on a Wind Turbine Test Rig
,
2014
.
[5]
Christophe Bérenguer,et al.
Method for computing efficient electrical indicators for offshore wind turbine monitoring
,
2014
.
[6]
Daming Lin,et al.
A review on machinery diagnostics and prognostics implementing condition-based maintenance
,
2006
.
[7]
Corinne Mailhes,et al.
AStrion strategy: from acquisition to diagnosis. Application to wind turbine monitoring
,
2016
.
[8]
Corinne Mailhes,et al.
Time-Frequency Tracking of Spectral Structures Estimated by a Data-Driven Method
,
2015,
IEEE Transactions on Industrial Electronics.
[9]
Corinne Mailhes,et al.
Consequences of non-respect of the Bedrosian theorem when demodulating
,
2013
.