Filtern
Dokumenttyp
- Vortrag (7) (entfernen)
Sprache
- Englisch (7)
Referierte Publikation
- nein (7)
Schlagworte
- Machine learning (7)
- Distributed fiber-optic sensors (2)
- Fiber optic sensors (2)
- Fiber optics sensors (2)
- Hydrogen (2)
- Pressure vessels (2)
- Structural health monitoring (2)
- Acoustic emission (1)
- Artificial neural networks (1)
- BOFDA (1)
Organisationseinheit der BAM
Eingeladener Vortrag
- nein (7)
Brillouin Optical Frequency Domain Analysis (BOFDA) is a powerful and well-established method for static distributed sensing of temperature and strain. Recently, we demonstrated a BOFDA system based on convolutional neural network which shortens the measurement time considerably. In this paper, we apply leave-one-out cross validation to evaluate the generalization performance and provide an unbiased and reliable machine learning model for a time-efficient BOFDA system.
In the last few years, the use of machine learning has emerged in the field of distributed fiber optic sensors as a promising approach to enhance their performance and provide new capabilities. In this study, we use machine learning for simultaneous measurements of temperature and humidity in polyimide (PI)-coated optical fibers based on Brillouin Brillouin optical frequency domain analysis (BOFDA). Different non-linear machine learning algorithms are employed, namely polynomial regression, decision trees and artificial neural networks (ANNs), and their discrimination performance is benchmarked against that of the conventional linear regression. The performance is evaluated using leave-one-out cross-validation to ensure that the models are reliable and able to generalize well on new data. We show that nonlinear machine learning algorithms outperform the conventional linear regression and thus could pave the way towards simultaneous cost-effective temperature and humidity distributed sensing, which has the potential to find attractive new applications in the field of civil and geotechnical engineering, from structural health monitoring of dikes and bridges to subsea cables and long pipelines corrosion detection.
Short presentation of the PhD project in machine learning based Brillouin distributed sensing. Machine learning can be used to enhance the performance of BOFDA and reduce considerably the measurement time. Apart from this, ML can also be used to extract more information from the Brillouin gain spectrum in order to render the temperature and strain discrimination possible
Brillouin Optical frequency domain analysis (BOFDA) is a well-known technique in distributed Fiber sensing for temperature and strain monitoring. In this study, we applied a signal post-processing method based on convolutional neural networks (CNNs) and we show that in comparison to conventional methods based on Lorentzian curve fitting (LCF), our CNN model is more robust against noise and can shorten the measurement time by more than 9 times. This will open the way for applications, where faster monitoring is needed.
We propose a method based on artificial neural networks to extract strain information from wavelength-scanning coherent optical time domain reflectometry (C-OTDR) data. Our neural network algorithm performs more than two orders of magnitude faster than the conventional approach.
This is due to the highly parallel evaluation of the neural networks on a GPU accelerated computer and the fact that conventional correlation and interpolation analysis needs many Iteration steps. This opens the way for real-time C-OTDR strain sensing because the neural Network strain predictions require less time than the measurements themselves. Real-time data Analysis enables long-term sensing e.g. in structural health monitoring, because the large amount of raw data does not have to be stored but can immediately be reduced to the strain data of interest.
Structural health monitoring of hydrogen pressure vessels using distributed fiber optic sensing
(2024)
We report on distributed fiber optic sensing-based monitoring of hydrogen composite overwrapped pressure vessels (COPV) to simultaneously increase the operational lifespan and mitigate maintenance costs. Our approach represents, to the best of our knowledge, the first application of distributed fiber optic sensing for COPV Type IV monitoring, where the sensing fibers are attached to the surface, rather than integrated into the composite material. Specifically, we attach an optical fiber of 50 m to the pressure vessel's surface, covering both the cylindrical and dome sections. We note that our fiber optic sensing technique relies on swept wavelength interferometry providing strain information along the entire length of the optical fiber with high spatial resolution even at the millimeter scale. When the vessel is pressurized, the sensing optical fiber shows a linear strain response to pressure at every position along the fiber. After thousands of load cycles, the vessel finally fails with the optical fiber detecting and precisely localizing the damage in the vessel’s blind dome area. Furthermore, we discuss the potential of state-of-the-art signal processing methods and machine learning for advancing predictive maintenance. This could reduce the number of regular inspections, mitigate premature maintenance costs, and simultaneously increase the vessel’s remaining safe service life. We believe that the structural health monitoring of hydrogen pressure vessels with fiber optic sensors can enhance trust in hydrogen technology contributing to the energy transition in the future.
Hydrogen pressure vessels are among the most essential components for reliable hydrogen technology. Under current regulations, a mostly conservative strategy is employed, restricting the usage time of hydrogen pressure vessels without providing information on the real remaining lifetime. During the service life, pressure vessels are inspected periodically. However, no established method that can provide continuous monitoring or information on the remaining safe service life of the vessel. In this paper, we propose a sensor network for Structural Health Monitoring (SHM) of hydrogen pressure vessels where data from all sensors are collected and centrally evaluated. Specifically, we integrate three different SHM sensing technologies namely Guided Wave ultrasonics (GW), Acoustic Emission testing (AT), and distributed Fiber Optic Sensing (FOS). This integrated approach offers significantly more information and could therefore enable a transition from costly and time-consuming periodic inspections to more efficient and modern predictive maintenance strategies, including Artificial Intelligence (AI)-based evaluation. This does not only have a positive effect on the operational costs but enhances safety through early identification of critical conditions in the overall system in real-time. We demonstrate an experimental set-up of a lifetime test where a Type IV Composite Overwrapped Pressure Vessel (COPV) is investigated under cyclic loading instrumented with AT, FOS, and GW methods. We acquired data from the sensor network until the pressure vessel failed due to material degradation. The data collected using the three different SHM sensor technologies is planned to be evaluated individually, using data fusion, and AI. In the future, we aim to integrate the measurement setup into a hydrogen refueling station with the data stream implemented into a digital signal processing chain and a digital twin.