Analytische Chemie
Filtern
Dokumenttyp
Sprache
- Englisch (19)
- Mehrsprachig (1)
Schlagworte
- Machine learning (5)
- Distributed acoustic sensing (4)
- Artificial neural networks (3)
- BOFDA (3)
- DAS (3)
- Distributed Fiber Optic Sensor (3)
- Geometric Phase (3)
- Brillouin distributed sensing (2)
- Coherent Heterodyne (2)
- Fiber optics sensors (2)
Organisationseinheit der BAM
Time-Efficient Convolutional Neural Network-Assisted Brillouin Optical Frequency Domain Analysis
(2021)
To our knowledge, this is the first report on a machine-learning-assisted Brillouin optical frequency domain analysis (BOFDA) for time-efficient temperature measurements. We propose a convolutional neural network (CNN)-based signal post-processing method that, compared to the conventional Lorentzian curve fitting approach, facilitates temperature extraction. Due to its robustness against noise, it can enhance the performance of the system. The CNN-assisted BOFDA is expected to shorten the measurement time by more than nine times and open the way for applications, where faster monitoring is essential.
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.
Solution doping of commercial polymethyl methacrylate (PMMA) polymer optical fibers (POFs) is presented as a novel approach for preparation of custom cladding-doped POFs (CD-POFs). The presented method is based on a solution-mediated diffusion of dopant molecules into the fiber cladding upon soaking of POFs in a methanol-dopant solution. The method was tested on three different commercial POFs using Rhodamine B as a fluorescent dopant. The dynamics of the diffusion process was studied in order to optimize the doping procedure in terms of selection of the most suitable POF, doping time and conditions. Using the optimized procedure, longer segment of fluorescent CD-POF was prepared and its performance was characterized. Fiber’s potential for sensing and illumination applications was demonstrated and discussed. The proposed method represents a simple and cheap way for fabrication of custom, short to medium length CD-POFs with various dopants.
Solution doping of off-the-shelf plastic optical fibers (POFs) represents rather simple and cheap way for preparing custom cladding-doped POFs (CD-POFs) with short to medium lengths. CD-POFs are especially attractive for environmental sensing applications, but might be of interest for illumination task as well. In this work, the proposed doping technique is tested with three different commercial low-cost polymethyl methacrylate (PMMA) POFs; Eska CK-40 and Eska GK-40 from Mitsubishi Rayon and Raytela PGU-FB1000 from Toray. The aim of the work is to aid the selection of the most suitable fiber yielding highest optical quality of prepared CD-POFs. Firstly, the optimal doping times are determined for the individual fiber types using short fiber samples. Secondly, longer 10 m CD-POFs are prepared from all tested fibersusing the optimized doping procedure. Finally, attenuation of pristine POFs and prepared CD-POFs is measured using optical time domain reflectometry in order to characterize the impacts of the doping on fiber optical properties. In addition, the importance of post-doping drying procedure for CD-POF optical performance is investigated as well. The results suggest that, although doping of all tested fibers is generally feasible, Eska CK-40 is the most suitable candidate with regard to the Doping efficiency and fiber post-doping performance.
Measurement and data recording systems are important parts of a holistic Structural Health Monitoring (SHM) system.
New sensor technologies such as fiber-optic sensors are often used; however, standards (or at least guidelines) are not yet available or internationally approved. This lack in standardization makes the acceptance of FOS technologies in complex SHM systems substantially difficult. A standard family for different FOS technologies is therefore being developed that should help to design SHM systems in an optimal way. International standardization activities take place in several standardization bodies such as IEC and ASTM, and within SHM societies such as ISHMII.
The paper reports on activities in standardization of fiber-optic sensors, on results already achieved, and on newly started projects. Combined activities of fiber sensor experts and SHM experts from Civil Engineering are presented. These contributions should help owners of structures as well as developers of sensors and monitoring systems to select effective and validated sensing technologies. Using these standards, both parties find recommendations how to proceed in development of SHM systems to evaluate the structural behavior based on e.g. standardized fiber optic sensors, and to derive necessary measures, e.g. the optimal maintenance strategy.
Geometric phase measured per beat period in a ϕ-OTDR based on coherent heterodyne detection is used to measure strain. Proposed method is robust to polarisation mismatch fading as a polarisation mismatch between interfering beams is not a hindrance to the measurement of the geometric phase. The Geometric phase is a function of the intensities of the interfering beams as well as the envelope of the beat signal. Its calculation does not require phase unwrapping and accordingly does not suffer the phase unwrapping errors. It is required to be equated with the traditionally measured phase by applying a scaling factor. The spatial resolution of the measured strain is reduced as it is calculated per beat period. Results are verified using a piezo-electric transducer inline a fiber-under-test.
A phase-sensitive optical time domain reflectometer based on coherent heterodyne detection of geometric phase in the beat signal of light, is reported for the first time to our knowledge. The use of the geometric phase to extract strain makes it immune to polarisation diversity fading. This is because a polarisation mismatch between the interfering beams is not a hindrance to its measurement. The geometric phase is calculated using the amplitude of the beat signal and individual beam intensities without any need for phase unwrapping. It is measured per beat period and can be equated with the traditionally measured dynamic phase with appropriate scaling. The results show that the system based on the geometric phase successfully measures strain, free from polarisation mismatch fading and phase unwrapping errors, providing a completely novel solution to these problems.
A systematic study of annealing behavior of drawn PMMA fibers was performed. Annealing Dynamics were investigated under different environmental conditions by fiber longitudinal shrinkage monitoring. The shrinkage process was found to follow a stretched exponential decay function revealing the heterogeneous nature of the underlying molecular dynamics. The complex dependence of the fiber shrinkage on initial degree of molecular alignment in the fiber, annealing time and temperature was investigated and interpreted. Moreover, humidity was shown to have a profound effect on the annealing process, which was not recognized previously. Annealing was also shown to have considerable effect on the fiber mechanical properties associated with the relaxation of molecular alignment in the fiber. The consequences of fiber annealing for the climatic stability of certain polymer optical fiber-based sensors are discussed, emphasizing the importance of fiber controlled pre-annealing with respect to the foreseeable
operating conditions.
We report, to our knowledge for the first time on simultaneous distributed temperature and strain sensing in a standard telecom optical fiber using a machine learning assisted Brillouin frequency domain analysis (BOFDA) system. The well-known temperature and strain cross-sensitivity problem is addressed by developing a BOFDA system with a high signal-to-noise ratio and applying machine learning. The spectrum consists of four highly resolved peaks, whose Brillouin frequency shifts are extracted and serve as features for the machine learning algorithms. The spectra result from a 450-m standard SMF-28 optical fiber, and particularly from a segment of 30 m. This fiber segment is coiled around a stretcher and placed in a climate chamber. The applied temperature and strain values range from 20 °C to 40 °C and from 0 µɛ to 1380 µɛ, respectively. The total measurement time to achieve a high SNR and resolve four peaks with a spatial resolution of 6 m is 16 min. To discriminate temperature and strain effects, simple frequentist and more sophisticated Bayesian-based algorithms are employed with the powerful Gaussian process regression (GPR) delivering the best performance in terms of temperature and strain errors, which are found to be 2 °C and 45 µɛ, respectively. These errors are calculated using leave-one-out cross-validation, so that an unbiased estimation of the sensor’s performance is provided.
This paper presents reported machine learning approaches in the field of Brillouin distributed fiber optic sensors (DFOSs). The increasing popularity of Brillouin DFOSs stems from their capability to continuously monitor temperature and strain along kilometer-long optical fibers, rendering them attractive for industrial applications, such as the structural health monitoring of large civil infrastructures and pipelines. In recent years, machine learning has been integrated into the Brillouin DFOS signal processing, resulting in fast and enhanced temperature, strain, and humidity measurements without increasing the system’s cost. Machine learning has also contributed to enhanced spatial resolution in Brillouin optical time domain analysis (BOTDA) systems and shorter measurement times in Brillouin optical frequency domain analysis (BOFDA) systems. This paper provides an overview of the applied machine learning methodologies in Brillouin DFOSs, as well as future perspectives in this area.