Monitoring von Turmstrukturen mit probabilistischen Methoden: Monitoring of Tower Structures with Probabilistic Methods

Research output: Chapter in book/report/conference proceedingContribution to book/anthologyResearch

View graph of relations

Details

Translated title of the contributionMethoden Monitoring of Tower Structures with Probabilistic Methods
Original languageGerman
Title of host publicationBaudynamik 2025
Pages135-146
Number of pages12
ISBN (electronic)9783181024478
Publication statusPublished - 2025

Publication series

NameVDI-Berichte
Number2447
Volume2025
ISSN (Print)0083-5560

Abstract

As the hub height of wind turbines increases, so do the demands for monitoring the structural integrity their tower structures. Vibration-based structural health monitoring often employs operational modal analysis methods, such as Bayesian Operational Modal Analysis (BAYOMA). This study investigates the impact of closely spaced modes of tower structures on identification uncertainties. It is shown that both identification and identification uncertainties are subject to operational influences. Consequently, heteroscedastic Gaussian Processes (GP) offer robust data normalization that accounts for this input-dependent variance. A newly developed probabilistic novelty metric combines the identification uncertainties from BAYOMA with the regression uncertainties from the GP for condition assessment. The described approach is tested on the tower structure of an operating wind turbine.

Cite this

Monitoring von Turmstrukturen mit probabilistischen Methoden: Monitoring of Tower Structures with Probabilistic Methods. / Jonscher, C.; Grießmann, T.; Rolfes, R.
Baudynamik 2025. 2025. p. 135-146 (VDI-Berichte; Vol. 2025, No. 2447).

Research output: Chapter in book/report/conference proceedingContribution to book/anthologyResearch

Download
@inbook{d63fa61fb7e940879c2bd0740e8471d7,
title = "Monitoring von Turmstrukturen mit probabilistischen Methoden: Monitoring of Tower Structures with Probabilistic Methods",
abstract = "As the hub height of wind turbines increases, so do the demands for monitoring the structural integrity their tower structures. Vibration-based structural health monitoring often employs operational modal analysis methods, such as Bayesian Operational Modal Analysis (BAYOMA). This study investigates the impact of closely spaced modes of tower structures on identification uncertainties. It is shown that both identification and identification uncertainties are subject to operational influences. Consequently, heteroscedastic Gaussian Processes (GP) offer robust data normalization that accounts for this input-dependent variance. A newly developed probabilistic novelty metric combines the identification uncertainties from BAYOMA with the regression uncertainties from the GP for condition assessment. The described approach is tested on the tower structure of an operating wind turbine.",
author = "C. Jonscher and T. Grie{\ss}mann and R. Rolfes",
note = "Publisher Copyright: {\textcopyright} 2025, VDI Verlag GMBH. All rights reserved.",
year = "2025",
doi = "10.51202/9783181024478-135",
language = "Deutsch",
isbn = "978-3-18-092447-2 ",
series = "VDI-Berichte",
number = "2447",
pages = "135--146",
booktitle = "Baudynamik 2025",

}

Download

TY - CHAP

T1 - Monitoring von Turmstrukturen mit probabilistischen Methoden

T2 - Monitoring of Tower Structures with Probabilistic Methods

AU - Jonscher, C.

AU - Grießmann, T.

AU - Rolfes, R.

N1 - Publisher Copyright: © 2025, VDI Verlag GMBH. All rights reserved.

PY - 2025

Y1 - 2025

N2 - As the hub height of wind turbines increases, so do the demands for monitoring the structural integrity their tower structures. Vibration-based structural health monitoring often employs operational modal analysis methods, such as Bayesian Operational Modal Analysis (BAYOMA). This study investigates the impact of closely spaced modes of tower structures on identification uncertainties. It is shown that both identification and identification uncertainties are subject to operational influences. Consequently, heteroscedastic Gaussian Processes (GP) offer robust data normalization that accounts for this input-dependent variance. A newly developed probabilistic novelty metric combines the identification uncertainties from BAYOMA with the regression uncertainties from the GP for condition assessment. The described approach is tested on the tower structure of an operating wind turbine.

AB - As the hub height of wind turbines increases, so do the demands for monitoring the structural integrity their tower structures. Vibration-based structural health monitoring often employs operational modal analysis methods, such as Bayesian Operational Modal Analysis (BAYOMA). This study investigates the impact of closely spaced modes of tower structures on identification uncertainties. It is shown that both identification and identification uncertainties are subject to operational influences. Consequently, heteroscedastic Gaussian Processes (GP) offer robust data normalization that accounts for this input-dependent variance. A newly developed probabilistic novelty metric combines the identification uncertainties from BAYOMA with the regression uncertainties from the GP for condition assessment. The described approach is tested on the tower structure of an operating wind turbine.

UR - http://www.scopus.com/inward/record.url?scp=105007200604&partnerID=8YFLogxK

U2 - 10.51202/9783181024478-135

DO - 10.51202/9783181024478-135

M3 - Beitrag in Buch/Sammelwerk

SN - 978-3-18-092447-2

T3 - VDI-Berichte

SP - 135

EP - 146

BT - Baudynamik 2025

ER -

By the same author(s)