Back to search

University of Patras

Prognostic methodologies for remaining useful life estimation of aerospace structures utilizing structural health monitoring technologies

Abstract

dc:description

In the present thesis the possibility of estimating of the Remaining Useful Life of composite aero-structures utilizing Structural Health Monitoring (SHM) data is investigated. More specifically, the aim is to develop a prognostic framework capable of monitoring the degradation of stiffened composite panels using advanced Health Indices and eventually estimate their Remaining Useful Life (RUL). The framework is built through extensive testing in single-stiffened panels and attempts to upscale the methodology to complex multi-stiffened panels and provide insight on the capabilities of SHM for such structures. In an era where concepts such as Prognostic and Health Management (PHM) and Condition Based Maintenance (CBM) are becoming increasingly popular methodologies that effectively monitor the real-time condition of structures are imperative. SHM technology provides a solution, where a capable sensor network collects the necessary data. To this direction, an extensive experimental campaign is launched, including compression fatigue experiments of different amplitudes and a couple of SHM technologies, i.e., FBG strain sensors and acoustic emission sensors, are used to monitor the real time degradation under variable loading conditions. RUL prognosis is the ultimate level of SHM, where conditional on the SHM data probabilistic estimates of the time to failure are aimed. The quality of the SHM data greatly affects the accuracy of the prognostic algorithms and potential features that can capture the structural degradation are sought after. To identify capable degradation features, an investigation is launched to create Health Indicators (HIs), that are able to capture the growing damage in variable loading conditions. By processing the raw SHM data, few HIs are identified that correlate to physical properties while some others are created with advanced processing methodologies in order to possess desirable attributes like monotonicity and prognosability. To further improve those HIs, an optimization scheme based on Genetic Algorithms (GA) is utilized. With the aid of machine learning or statistical algorithms, the discovered features are used to estimate the RUL. Despite the variable operational conditions (i.e. loads), the proposed HI was able to improve the accuracy of the predictions which was a result of the improved monotonic and prognosable behaviors we requested in our optimization problem. The predicted RUL shows promising accuracy especially near the end of life, which is further solidified by the decreasing of the confidence intervals width. In the final part of my thesis, an upscaling attempt is presented. More specifically, the methodologies developed in single stiffened panels are investigated for their applicability at multi-stiffened panels. A challenging fatigue experimental campaign is launched with variable amplitude fatigue tests where distributed fiber optical strain sensors are used to monitor their degradation. Methodologies for data processing and HI creation, discovered and validated on single stiffened panels, are applied to the multi-stiffened ones. Machine learning algorithms are trained with the HIs from lower hierarchically structures i.e. the single-stiffened panels and are used to predict the RUL of multi-stiffened panels with promising accuracy demonstrating the capabilities of the proposed methodologies.

Degree

thesis:*
Grantor dc:publisher
University of Patras
Year dc:date
2023

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Galanopoulos, Georgios
  • Γαλανόπουλος, Γεώργιος

Subjects

dc:subject × 32

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/55142

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Galanopoulos, Georgios; Γαλανόπουλος, Γεώργιος. Prognostic methodologies for remaining useful life estimation of aerospace structures utilizing structural health monitoring technologies. University of Patras, 2023. http://hdl.handle.net/10442/hedi/55142