Detection of particle contamination and lubrication outage in journal bearings in wind turbine gearboxes using surface acoustic wave measurements and machine learning

被引:0
|
作者
Decker, Thomas [1 ]
Jacobs, Georg [1 ]
Raddatz, Malte [1 ]
Roeder, Julian [1 ]
Betscher, Jonas [1 ]
Arneth, Philipp [2 ]
机构
[1] Rhein Westfal TH Aachen, Chair Wind Power Drives, Campus Blvd 61, D-52074 Aachen, Germany
[2] BestSens AG, Jean Paul Weg 2, D-93489 Niederfullbach, Germany
来源
关键词
D O I
10.1007/s10010-025-00784-4
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
Journal bearings are used more and more in wind turbine (WT) gearboxes. Compared to rolling element bearings they are advantageous in terms of power density and reliability. Despite their reliability and theoretically unlimited fatigue life, journal bearings can be damaged by particularly critical operating conditions that do not represent normal WT operation. As journal bearing damage can occur very suddenly in the worst case, continuous monitoring of the bearing's condition is advisable. Particle contamination in the lubricant and an outage of the oil supply can be particularly harmful to the bearing. Condition monitoring systems (CMS) have the potential to detect such critical operating conditions in journal bearings before damage occurs. A detection of these conditions is crucial for preventing bearing damage and thus gearbox failure which results in turbine downtime and yield loss. If failures of journal bearings in WT gearboxes can be avoided through the use of CMS this, in the long term, has the potential to reduce maintenance and repair costs in the field application.In this work the novel surface acoustic wave (SAW) measurement method is presented for the detection of particle contamination and lubrication outage. The SAW method is advantageous compared to conventional monitoring methods such as vibration measurements, as it is based on measuring the propagation behavior of actively introduced SAW into the bearing. This makes the method particularly robust against disturbing noise. For the evaluation of the signals and the detection of the aforementioned operational anomalies a machine learning approach is used. The latter is implemented such that an online monitoring can be performed with only a short latency between data input and evaluation.The presented method was validated on a component test rig for journal bearings. For the experiments the SAW measurement was implemented into the test bearings. In the test campaign, the anomalies were actively induced and the bearing behavior observed over time. This work provides insight into the signals measured during the occurrence of operational anomalies and proves that a lubrication outage and particle contamination can be detected using SAW. Gleitlager finden zunehmend in Getrieben von Windenergieanlagen (WEA) Anwendung. Im Vergleich zu W & auml;lzlagern bieten sie Vorteile hinsichtlich Leistungsdichte und Zuverl & auml;ssigkeit. Trotz ihrer hohen Zuverl & auml;ssigkeit k & ouml;nnen Gleitlager durch kritische Betriebsbedingungen, die vom normalen Betrieb abweichen, besch & auml;digt werden. Da Sch & auml;den an Gleitlagern oft pl & ouml;tzlich auftreten, ist eine kontinuierliche & Uuml;berwachung des Lagers ratsam. Partikelkontamination im Schmiermittel und ein Ausfall der & Ouml;lversorgung sind besonders sch & auml;dlich f & uuml;r das Lager. Zustands & uuml;berwachungssysteme (CMS) k & ouml;nnen solche kritischen Betriebsbedingungen fr & uuml;hzeitig erkennen und helfen, Lagersch & auml;den und damit Getriebesch & auml;den zu verhindern. Durch den Einsatz von CMS lassen sich langfristig Wartungs- und Reparaturkosten senken.In dieser Arbeit wird eine Messmethode mit akustischen Oberfl & auml;chenwellen (Surface Acoustic Waves, SAW) zur Erkennung von Partikelkontamination und & Ouml;lversorgungs-Ausf & auml;llen vorgestellt. Die SAW-Methode bietet im Vergleich zu herk & ouml;mmlichen & Uuml;berwachungsmethoden wie Vibrationsmessungen Vorteile, da sie auf dem Ausbreitungsverhalten aktiv eingef & uuml;hrter SAW im Lager basiert. Dadurch ist die Methode robust gegen & uuml;ber St & ouml;reinfl & uuml;ssen. Zur Auswertung der Signale und zur Erkennung von Betriebsanomalien wird ein maschinelles Lernverfahren eingesetzt, das eine Online-& Uuml;berwachung mit minimaler Verz & ouml;gerung zwischen Dateneingabe und Auswertung erm & ouml;glicht.Die Methode wurde an einem Komponentenpr & uuml;fstand f & uuml;r Gleitlager mit integrierter SAW-Messung validiert. In den Experimenten wurden die Anomalien aktiv induziert und das Lagerverhalten & uuml;ber einen bestimmten Zeitraum beobachtet. Diese Arbeit zeigt auf, dass ein Ausfall der & Ouml;lversorgung sowie Partikelkontamination mithilfe der SAW-Technologie erkannt werden k & ouml;nnen.
引用
收藏
页数:13
相关论文
共 36 条
  • [21] Crack detection and localization on wind turbine blade using machine learning algorithms: A data mining approach
    Joshuva A.
    Sugumaran V.
    SDHM Structural Durability and Health Monitoring, 2019, 13 (02): : 181 - 203
  • [22] Specific Detection of BTEX Contamination in Water Using a π-Hole-Catching Surface Acoustic Wave Sensor
    Friedt, Jean-Michel
    Luzet, Vincent
    Soumann, Valerie
    Nief, Nathalie
    Segues, Bertrand
    Pucheu, Gilles
    Dehez, Jean-Sebastien
    Ordonez-Varela, John-Richard
    Cherioux, Frederic
    ACS OMEGA, 2025, 10 (11): : 10903 - 10910
  • [23] A Novel Method for Smart Fire Detection Using Acoustic Measurements and Machine Learning: Proof of Concept
    John Martinsson
    Marcus Runefors
    Håkan Frantzich
    Dag Glebe
    Margaret McNamee
    Olof Mogren
    Fire Technology, 2022, 58 : 3385 - 3403
  • [24] A Novel Method for Smart Fire Detection Using Acoustic Measurements and Machine Learning: Proof of Concept
    Martinsson, John
    Runefors, Marcus
    Frantzich, Hakan
    Glebe, Dag
    McNamee, Margaret
    Mogren, Olof
    FIRE TECHNOLOGY, 2022, 58 (06) : 3385 - 3403
  • [25] Machine-Learning-Based Detection of Aerodynamic Disturbances Using Surface Pressure Measurements
    Hou, Wei
    Darakananda, Darwin
    Eldredge, Jeff D.
    AIAA JOURNAL, 2019, 57 (12) : 5079 - 5093
  • [26] Anomaly detection using a SCADA feature extractor and machine learning to detect lightning damage on wind turbine blades
    Matsui, Takuto
    Yamamoto, Kazuo
    Ogata, Jun
    IEEJ TRANSACTIONS ON ELECTRICAL AND ELECTRONIC ENGINEERING, 2022, 17 (06) : 945 - 951
  • [27] Predictive maintenance of abnormal wind turbine events by using machine learning based on condition monitoring for anomaly detection
    Huan Chen
    Jyh-Yih Hsu
    Jia-You Hsieh
    Hsin-Yao Hsu
    Chia-Hao Chang
    Yu-Ju Lin
    Journal of Mechanical Science and Technology, 2021, 35 : 5323 - 5333
  • [28] Predictive maintenance of abnormal wind turbine events by using machine learning based on condition monitoring for anomaly detection
    Chen, Huan
    Hsu, Jyh-Yih
    Hsieh, Jia-You
    Hsu, Hsin-Yao
    Chang, Chia-Hao
    Lin, Yu-Ju
    JOURNAL OF MECHANICAL SCIENCE AND TECHNOLOGY, 2021, 35 (12) : 5323 - 5333
  • [29] Feature detection in guided wave ultrasound measurements using simulated spectrograms and generative machine learning
    Setshedi, Isaac I.
    Wilke, Daniel N.
    Loveday, Philip W.
    NDT & E INTERNATIONAL, 2024, 143
  • [30] Achieving robust damage mode identification of adhesive composite joints for wind turbine blade using acoustic emission and machine learning
    Xu, D.
    Liu, P. F.
    Chen, Z. P.
    Leng, J. X.
    Jiao, L.
    COMPOSITE STRUCTURES, 2020, 236