Robust SHM

Redundancy approach with different sensor integration levels for long life monitoring systems

authored by: Jan Hauke Bartels, Thomas Potthast, Sören Möller, Tanja Grießmann, Raimund Rolfes, Michael Beer, Steffen Marx
Abstract: Structural health monitoring (SHM) techniques use a variety of sensors, such as displacement, strain, and acceleration sensors, to assess the current condition of engineering structures that are designed to last for decades. Over time, structures can experience degradation-related damage, while the monitoring systems themselves can age and degrade, becoming less reliable. This aging can lead to sensor malfunctions that produce plausible but incorrect data, leading to misinterpretations of structural integrity and potentially catastrophic failures. Therefore, it is critical to distinguish between sensor anomalies and structural anomalies to ensure robust SHM throughout the life cycle of the structure. To address this issue, this study introduces a two-step redundancy approach. First, strain gauges were aged in a climate chamber in laboratory experiments to determine the time-variant behavior of the measurement system. Measurement drift and physical gain were identified as significant changes in sensor performance. Second, the redundancy approach using correlation analysis and Gaussian process regression was used to effectively detect and isolate these sensor anomalies. The method successfully distinguished between sensor and structural anomalies and proved to be robust in various scenarios. Further research is suggested to explore the reliability of this approach under conditions where structural and sensor anomalies coincide. This study enhances the robustness of SHM systems and supports reliable assessment of structural health over their lifetime.
Organisation(s): Institute for Risk and Reliability
Institute of Structural Analysis
External Organisation(s): Technische Universität Dresden
University of Liverpool
Tongji University
Type: Paper
Publication date: 01.07.2024
Publication status: Published
Peer reviewed: Yes
ASJC Scopus subject areas: Computer Science Applications, Health Information Management
Electronic version(s): https://doi.org/10.58286/29699 (Access: Open)

BibTeX

@conference{ceea319699a8431980d79a38d0a5b1f3,
title = "Robust SHM: Redundancy approach with different sensor integration levels for long life monitoring systems",
abstract = "Structural health monitoring (SHM) techniques use a variety of sensors, such as displacement, strain, and acceleration sensors, to assess the current condition of engineering structures that are designed to last for decades. Over time, structures can experience degradation-related damage, while the monitoring systems themselves can age and degrade, becoming less reliable. This aging can lead to sensor malfunctions that produce plausible but incorrect data, leading to misinterpretations of structural integrity and potentially catastrophic failures. Therefore, it is critical to distinguish between sensor anomalies and structural anomalies to ensure robust SHM throughout the life cycle of the structure. To address this issue, this study introduces a two-step redundancy approach. First, strain gauges were aged in a climate chamber in laboratory experiments to determine the time-variant behavior of the measurement system. Measurement drift and physical gain were identified as significant changes in sensor performance. Second, the redundancy approach using correlation analysis and Gaussian process regression was used to effectively detect and isolate these sensor anomalies. The method successfully distinguished between sensor and structural anomalies and proved to be robust in various scenarios. Further research is suggested to explore the reliability of this approach under conditions where structural and sensor anomalies coincide. This study enhances the robustness of SHM systems and supports reliable assessment of structural health over their lifetime.",
keywords = "Acceleratometers, Analytical redundancy, Correlation analysis, Fault detection, Fault diagnosis, Fault isolation, Gaussian process regression, Hardware redundancy, Sensor aging, Sensor integration level, Strain gauges, Structural health monitoring",
author = "Bartels, {Jan Hauke} and Thomas Potthast and S{\"o}ren M{\"o}ller and Tanja Grie{\ss}mann and Raimund Rolfes and Michael Beer and Steffen Marx",
note = "Publisher Copyright: {\textcopyright} 2024 11th European Workshop on Structural Health Monitoring, EWSHM 2024. All rights reserved.; 11th European Workshop on Structural Health Monitoring, EWSHM 2024 ; Conference date: 10-06-2024 Through 13-06-2024",
year = "2024",
month = jul,
day = "1",
doi = "10.58286/29699",
language = "English",
}