This PDF file contains the front matter associated with SPIE Proceedings Volume 13467, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

In Brillouin optical time domain analysis (BOTDA), ensuring both high measurement accuracy and computational efficiency is critical for reliable and scalable infrastructure monitoring. Measurement accuracy is influenced by factors such as noise level, frequency scanning resolution, pump pulse power and pulse width, as well as the method used to estimate the Brillouin frequency shift (BFS) which represents the frequency of peak of the Brillouin gain spectrum (BGS). In large-scale fiber deployments, computational efficiency becomes equally essential due to the high data volume and the time cost associated with signal averaging for noise suppression. This work presents a comparative evaluation of several BGS peak estimation methods aimed at maximizing both accuracy and speed. One approach employs a Euclidean distance-based technique applied to a reduced feature space generated through principal component analysis (PCA) of the BGS data. By compressing the input space to five principal components, this method achieves over 100 times faster processing speed than conventional Lorentzian curve fitting (LCF), while preserving or even improving accuracy depending on interrogation parameters. Building on this, we implement a probabilistic deep neural network (PDNN) trained on the same PCA-reduced data. The PDNN not only predicts the Brillouin frequency shift (BFS) but also estimates the variance, allowing direct computation of standard deviation and confidence intervals. Experimental validation was carried out by measuring temperature changes at the end of a 2 km fiber, with controlled heating to 35 °C and 45 °C. Our results confirm that both PCA-Euclidean and PCA-PDNN approaches maintain measurement errors below 0.8 °C. Moreover, PDNN provides robust BFS uncertainty quantification, yielding a 99% confidence interval for the measured BFS of ±1.935 MHz, and a standard deviation of ±0.75 MHz. This combination of high speed, accuracy, and statistical insight demonstrates the potential of PCA-PDNN models as a powerful solution for real-time, long-range BOTDA sensing in modern infrastructure applications.

Ultrasonic guided wave (UGW) techniques offer low energy attenuation and large coverage capabilities for nondestructive evaluation (NDE) and structural health monitoring (SHM), by the cost of traditionally lower sensitivity to smaller artifacts. On cylindrical structures, historical emphasis has predominantly been given to axisymmetric waves due to less complex interpretation. Flexural guided wave-based techniques are an alternative to the historically predominant axisymmetric modes, offering the important advantages of requiring only partial surface load to be performed, and multiple sensitivity characteristics, at the cost of need more sophisticated data analysis. This study seeks to leverage distributed and quasi-distributed optical fiber sensors as the elastic wave receivers to make use of the spatial-temporal capability of these sensors to overcome the complexity of multiple propagating guided wave modes and selected weld defects. A finite element model of an 8-ft long steel pipe, where flexural modes are generated using different PZT array configurations at the end of the pipe is used for the study case. The work investigates the sensitivity of different GW excitation methods and using distributed and quasi-distributed optical fiber sensors (DOFS and Q-DOFS) for dynamic strain sensing to take advantage of the spatial resolution uniquely provided by the sensor system to identify damages and welds characteristics. A simple convolutional neural network (CNN) with minimal depth and parameters is used to evaluate damage type to assess the impact on data quality, quantity and the sensing effectiveness. Overall, this comprehensive approach aims to advance the understanding of flexural guided wave propagation as a long range NDE method and its sensitivity to damage and weld characterization, a challenging task for these waves using traditional equipment.

Optical fiber sensors have become indispensable in a wide range of applications, with Fabry-Perot interferometers (FPIs) standing out for their compactness and versatility in sensor design. Microwave photonics (MWP) techniques offer enhanced performance and flexibility for advancing optical sensor interrogation methods. This paper presents a novel MWP-based interrogation technique, leveraging direct phase measurement for short-cavity FPI sensors. The method utilizes the phase response of the FPI sensor within an MWP-assisted radio frequency single bandpass filter, delivering improved sensitivity and dynamic sensing capabilities compared to conventional approaches. Experimental results confirm the effectiveness of the proposed technique for measuring both static and dynamic strains. The system achieves a high sensitivity of 0.00217 rad/με and a resolution better than 100 nε at an operating frequency of 247.0 MHz.

A graphene metasurface, which could function as a “spectrometer-on-a-chip” for early-stage integrated optics devices, is theoretically and experimentally analyzed for detection of narrowband long-wave infrared (LWIR) emissions. Graphene films are electrically reconfigurable and lie atop a Fabry-Perot cavity, with plasmonic nanoparticles providing field enhancement. The narrowband absorption/emission of the graphene can be scanned over the long-wave infrared (LWIR) regime from 4 to 14 um by modulating the local chemical potential with a voltage. Efficient narrowband measurements of the emission spectrum of room-temperature objects can be made, or quantify adsorbed or fluorescent molecules. Such a portable and inexpensive integrated optics and optoelectronic device could electrical system health (e.g., utility system), energy infrastructure, carbon or methane management systems, or military systems. Graphene or spraycoated devices could be conformable on different, non-planar shapes. Visible color for these portable spectrometers is supplied by spraycoated biopigment Xanthommatin. Experimental results from coated samples will be reported.

Evanescent-field optical fiber sensors offer a wide range of benefits for gas and chemical sensing applications, including low cost per sensor node, stability under conditions that would be unfriendly for electronic sensor packages (e.g., high temperature, chemically harsh conditions, electromagnetic interference), and a small “footprint” for space-constrained applications. Optical fiber sensors also offer multiple unique capabilities, such as multi-wavelength interrogation via wavelength multiplexing and compatibility with spatially distributed interrogation techniques for multi-point detection. This approach relies upon the development of materials that demonstrate (1) high sensitivity, (2) selectivity, (3) fast response, and (4) long-term stability under relevant operating conditions. In this work, machine learning tools will be demonstrated to improve the performance in these areas, using simulated sensor data based on a simple, idealized sensor system. The application of recurrent neural networks in the form of a long short-term memory (LSTM) model will be applied to simulated time-series sensor data to demonstrate improvement in selectivity and effective response time. Multi-parameter sensing applications relevant to energy sector applications will be discussed.

Ensuring the safety, integrity, and operational efficiency of underground product pipelines is vital for maintaining the nation’s critical infrastructure. Monitoring parameters such as hoop strain, pressure, and acoustic vibrations is key to detecting potential leaks, intrusions, or structural issues. Distributed optical fiber sensor (DOFS) systems provide a compelling solution for continuous, real-time monitoring over long distances. This paper details the development and pilot-scale implementation of DOFS systems for underground pipeline monitoring, evolving from a proof-of-concept stage. Multiple custom-designed DOFS interrogator units—such as optical frequency-domain reflectometry (OFDR), Brillouin optical time-domain analysis (BOTDA), and multimodal interferometer-based fiber acoustic sensor systems were tested to measure the key parameters, such as hoop strain, pipe pressure, surrounding soil temperature, and acoustic vibrations. The underground product pipeline’s outer diameter is 30 inches, the wall thickness is 1.28 inches, and 3 feet deep from the surface. The fiber deployment strategies and sensing data acquisition methods for these systems are discussed. The results demonstrate the effectiveness of DOFS in detecting hoop strain, temperature changes, and acoustic vibrations, showcasing their potential for real-time monitoring and enhancing pipeline safety. These findings from pilot-scale testing offer valuable insights into advancing pipeline monitoring technologies and improving the reliability of underground pipeline systems.

Deep learning has become a vital tool in ultrasonic testing for defect characterization, thanks to its automation and precision. However, in nondestructive evaluation (NDE), limited availability of experimental data presents a significant challenge. To address this, we simulated training datasets, which have proven effective. This paper presents an AI-driven approach for detecting pipeline damage using acoustic fiber sensors and guided waves, focusing on integrating experimental and simulated data through domain adaptation. Detecting damage in pipelines is essential for maintaining structural integrity, yet traditional methods often struggle due to scarcity of real-world data. To overcome this, we use finite element simulations to generate guided wave signals, which serve as the source domain. Concurrently, experimental signals from acoustic fiber sensors are treated as the target domain. We apply advanced domain adaptation techniques to align the feature spaces of both simulated and experimental data, ensuring accurate defect detection. A convolutional neural network (CNN) is employed to extract features that are consistent across both domains, thereby enhancing the reliability of damage identification even with limited experimental data. Our approach aims to improve classification accuracy, offering a robust solution for real-time pipeline monitoring. By effectively combining simulated and experimental data to enhance and expedite training of AI-based classification frameworks, we provide a comprehensive method for pipeline damage detection, ensuring structural health and safety.

Natural gas pipelines are critical infrastructure assets that require robust and continuous health monitoring to ensure safe and reliable operations. This paper presents an intelligent monitoring approach using fiber optic acoustic sensors based on a hybrid dataset comprising both real third-party intrusion and simulated guided wave signals. A 16-ft pipeline with inherent defects served as the test bed, where guided wave propagation experiments were conducted using a collar exciter at 32 kHz and a specialized single mode-multi mode-single mode (SMS) fiber optic acoustic sensor to detect signal responses. In parallel, third-party intrusion events were artificially generated via a speaker simulating various acoustic events, further broadening the dataset. The resulting data consisted of approximately 20 unique events, each played for 10 seconds twice. A physics-based guided wave simulation model was developed to replicate defects with varying dimensions and locations, producing a suite of virtual signals. Combining the experimentally measured and simulated signals yielded a robust hybrid dataset for advanced signal classification and regression. A convolutional neural network (CNN) was employed to classify events into physical defects versus non-physical intrusion signatures. Subsequently, for the classified physical defect signals, an XGBoost regressor was utilized to estimate defect variables (e.g., depth, length) and their positions within the pipeline. Results indicate that the proposed pipeline monitoring framework, enabled by fiber optic acoustic sensing and machine learning models, offers high accuracy in detecting and characterizing defects. This integrated approach can thus serve as an effective method for predictive maintenance, minimizing downtime and enhancing safety in natural gas pipeline operations.

Negative hydraulic pressure measurement finds broad applications across various industries and sciences, including biomedicine, food and beverage, liquid storage and transportation, as well as geotechnical engineering and agriculture. This talk introduces, for the first time, a novel optical fiber interferometric sensor designed for direct local measurement of negative hydraulic pressure. The sensor incorporates a dual-micro-cavity structure consisting of a hollow airgap Fabry-Perot interferometer (FPI) cavity and a liquid reservoir, separated by a thin silica diaphragm. A partially fusion-collapsed photonic crystal fiber serves as the entry channel to the liquid reservoir, facilitating the transfer of negative hydraulic pressures to the liquid reservoir. This pressure transfer causes slight deflection of the silica diaphragm, which is precisely measured by the FPI. The prototype device demonstrates the capability to measure both positive and negative pressures across a wide range of pressures from -1000 kPa to 1000 kPa with a resolution of 0.8 kPa.

Metal storage canisters play an important role in the nuclear industry for the safe storage and transport of spent nuclear fuel and radioactive waste. These metal canisters are typically constructed from stainless steel, are structurally robust and are hermetically sealed and welded to ensure effective radiation shielding. They are widely used in dry cask storage systems, with thick reinforced concrete layers for additional protection and radiation shielding. The structural integrity of these canisters is crucial for preventing leaks, ensuring containment, and complying with stringent nuclear regulatory standards throughout their service life. This work presents a real-time nondestructive evaluation (NDE) of stainless-steel canisters using fiber-optic acoustic sensing system employing Lamb-wave excitation. The study utilizes a stainless-steel Partial Canister Mockup (PCM) with simulated flaws including circumferential and axial seam-welds. The dispersion characteristics of the guided Lamb waves and resulting acoustic profiles are investigated experimentally and numerically, achieving good consistency in terms of propagation and acoustic characteristics. Through excitation and monitoring of the fundamental anti-symmetric (A₀) wave-mode, we successfully demonstrate the detection of polished circumferential weld in the PCM. The experimental findings are validated using numerical simulations, showing strong agreement in wave arrival times and echo signatures. The fiber-optic sensing approach enables long-range, high-resolution monitoring in remote or inaccessible environments, providing a scalable solution for ensuring the long-term integrity and safety of spent nuclear fuel canisters.

With an increased interest in subsurface gas storage technology for various energy applications, monitoring wellbore structural stability and subsurface geochemistry has become more pressing, and pH is a key parameter to measure. As high pressure and elevated temperatures in subsurface conditions are comparatively harsh relative to that expected for most standard pH sensor designs, any pH monitoring hardware must be designed for extended exposure to high pressures and temperatures. We previously reported that an optical fiber pH sensor functionalized with a calcined polyethylenimine coating had shown some promise as a high temperature and pressure pH sensor but with some drifting when operating for longer than 8 hours. In this paper, we investigated the coating composition and potential cause of the drifting and improved the stability of the prepared coating to minimize sensor drift under simulated wellbore conditions. Scanning electron microscopy (SEM) had previously shown moderate cracking at high pressures over short tests. By applying x-ray photoelectron spectroscopy (XPS) to characterize the sensor coating before and after one week of testing in an H2/CH4 gas blend at 80 °C and 900 psi, a compositional change in the coating was observable, which may indicate susceptibility to alteration by subsurface gas storage conditions. Non-reducing (CH4, N2) environments were also tested, and confirmed that both temperature and pressure were also contributing to the drift. As pH response showed an equilibration after initial acid and base response, we added a pre-treatment step of acid and base cycle to investigate the effect on stability and drifting.

Fiber Bragg grating (FBG) sensors have been widely applied for the collection of ultrasonic waves for structural health monitoring applications. One of the major challenges to using these systems for ultrasound detection is that the sensitivity of bonded FBG sensors to the low-amplitude waves is low. In this paper, we demonstrate strategies to increase the sensitivity of the FBG sensor to ultrasonic waves by converting the ultrasonic wave to a longitudinal acoustic mode in the optical fiber itself. The FBG then directly interacts with the longitudinal mode, to which its sensitivity is much higher. We address the coupling of the acoustic wave from the structure to the optical fiber and methods to tune this coupling for directional sensitivity.

This study presents a comparative evaluation of polymer-magnetic nanoparticle (MNP) and ferrofluid-coated multimode interferometric (MMI) fiber sensors for magnetic field sensing under AC field conditions. A Single Mode–No Core–Single Mode (SNS) fiber structure was fabricated and coated with a polyvinyl alcohol (PVA)– Fe₃O₄ nanocomposite via dip-coating to form a solid-state polymer-MNP sensor. The optical responses of both polymer and ferrofluid-coated sensors were characterized under sinusoidal AC fields at 60Hz and 300Hz, across discrete field strengths ranging from 50 to 700G. Under AC excitation, the polymer-MNP sensor exhibited a consistent second harmonic response, with a dominant optical signal at twice the frequency of the applied magnetic field. In contrast, the ferrofluid sensor displayed a combination of the fundamental and second harmonic components, with additional higher-order harmonics becoming more prominent at stronger fields. Fourier analysis and signal-to-noise ratio (SNR) evaluation confirmed that signal clarity improved with increasing magnetic field strength for both sensors. These results highlight the field-dependent optical response characteristics of each sensor type and demonstrate the potential viability of polymer-MNP-coated fiber sensors as compact, integrable platforms for dynamic magnetic field monitoring in power electronics and grid systems.

To ensure reliable operation and to prevent dangerous and costly failures, power transformers benefit from state-of- health (SOH) monitoring. A challenge with widespread deployment of SOH sensing and monitoring systems is the combination of technical challenge and economic cost. Fiber optic sensors are compact and can allow for in-situ measurement and fine-area resolution. This work focuses on integrating multiple evanescent-wave-based optical fiber sensors together into a compact interrogation unit for sequential sensing of hydrogen (H₂) gas, a well-known critical gas species in oil-immersed transformer health assessment. Sensing of 5% H₂ in a nitrogen (N₂) environment is performed within a custom gas chamber at room temperature. Opto-electrical conversion of sensor signals is subsequently performed within the interrogator utilizing low-cost components. The interrogator is also outfitted with wireless communication capability to transmit information over long distances with commercially available Radio Frequency (RF) wireless communication hardware. This work presents validation of rapid optoelectronic signal conversion and sequential data collection of two H₂ sensors and a self-referencing channel along with comprehensive data processing, resulting in a 3σ statistically significant limit-of-detection on the order of 500ppm of H₂ with a signal-to-noise ratio greater than 300.

Hydrogen has substantially different combustion properties from natural gas and can dramatically affect the operation of combustion systems. So to achieve operational flexibility in combustion equipment which may use blends of natural gas with hydrogen, fast analysis of the variation of gas composition is important. The National Energy Technology Laboratory’s (NETL) Raman gas analyzer (RGA) is a prototype field instrument for rapidly determining the composition of a gaseous mixture for real-time process control. The RGA uses a low power laser at visible wavelength and a reflective hollow waveguide as a sample gas cell for more intense Raman signals and fast sampling rates. The RGA unit has been calibrated with various pure gas species of interest including natural gas species and hydrogen, nitrogen, and oxygen. It has been characterized with varying hydrogen gas blends such as methane/hydrogen, ethane/hydrogen, and propane/hydrogen and natural gas model blends. In this paper, the measurement time and measurement accuracy of the composition of these gas blends at moderate pressures will be discussed. Demonstration of the RGA technology used to determine the composition of hydrogen blends quickly and continuously through laboratory and field tests, contributes to providing tools to ensure the safe and effective transport of hydrogen blends through existing natural gas and new gas pipelines.

This paper explores the reduction of background noise in a photoacoustic spectroscopy (PAS) system designed for trace methane measurements, specifically aimed at monitoring Undocumented orphaned wells (UOW). To reduce ambient noise and achieve robust field measurements, the resonant PAS sensor employs dual microphones for differential acoustic signal detection, significantly enhancing signal sensitivity and effectively suppressing flow-induced noise. The light source of the system is designed to achieve multiple reflections using two mirrors, thereby amplifying the signal tenfold after twelve reflections. Experimental results confirm the exceptional performance of the proposed system, which achieves a minimum detectable limit of 40 ppb with an averaging time of 100 seconds. The methane leak measurements were successfully carried out on-site of an abandoned natural gas well in Hillman State Park in the western Pennsylvania.

We fabricated methacrylate-based polymer waveguides and developed them into optical sensors to detect solvents and VOCs. When an analyte interacts with the polymer waveguide there is a dynamic gain in optical power due to the absorption of analyte molecules by the polymer which results in the swelling of the polymer matrix and gain in the optical power. This optical gain does not depend on the refractive index of the analyte rather it depends on the diffusion capacity of the analyte molecule in the polymer matrix, and hence using our waveguide we can differentiate among different analytes even if they have a similar refractive index which gives us an edge over the other optical sensors. Further, these waveguides were converted into temperature sensors to detect ambient temperatures, and stable results were observed by embedding gold nanoparticles inside the polymer matrix.

We investigate self-segregation of dopants in a crystal matrix within a single-crystal (SC) fiber. SC fiber is coated with dopant materials via a sol-gel dip coating method and used as the host material during laserheated pedestal growth (LHPG). Due to differences in self-segregation coefficients, dopants are pulled toward the center or periphery of the fiber via convection forces induced by the temperature gradient in the molten zone. This results in a SC fiber with a dopant-rich core and reduced-dopant cladding layer (or vice-versa). Here, we report on self-segregation of Cr and Ti dopants within a sapphire matrix induced during LHPG. Variations of different experimental parameters are investigated, including fiber pull speeds, laser power, and whether growth takes place within a reducing environment. The cross-sectional dopant concentrations are measured via electron-probe micro-analysis. The cross-sectional constituent profiles are characterized as a function of varying growth parameters to identify the conditions that result in the best quality fiber for distributed sensor applications.

Recent reports show hollow-core fibers (HCF) exceeding their silica counterpart in terms of transmission losses across UV to IR wavelengths. At shorter wavelengths, HCFs are of particular interest where solid core silica fibers suffer from scattering losses and light-induced solarization effect. In this work, a modified Particle Swarm Optimization (PSO) technique is employed to optimize a 12-capillary fiber design for minimal attenuation at 400nm. The PSO routine focuses on four coupled variables, enabling efficient optimization within the otherwise extensive parameter space for HCFs. The final design achieves a fundamental mode loss of 0.02 dB/km with 30μm mode field diameter and exhibits low surface scattering loss at 400nm of less than 0.001 dB/km. Such fiber design can facilitate high power delivery, and the realization of multi-core imaging HCFs with high pixel densities by utilizing the fiber structure as a unit cell for individual cores.

Halide fibers and waveguides, with transmission spanning approximately 1 micron to over 20 microns, have traditionally been used in midwave infrared (MWIR) and longwave infrared (LWIR) applications. The inclusion of silver halides in these optical waveguides renders them photosensitive. In this paper, we review the use of broadband infrared waveguides for sensor applications, fabricated using molten-core manufacturing technique. We explore the mechanisms behind photosensitive transformations and the potential applications of photoinduced structures in broadband infrared materials.

Distributed acoustic sensing (DAS) systems detect vibrations along optical fibers by analyzing Rayleigh backscattered light. The performance of DAS systems can be significantly impacted by Rayleigh backscattering generated from the sensing fiber. Standard telecom-grade single mode (SMF) fibers as well as Enhanced Scattering Fibers (ESFs) are used in DAS systems for various applications. In this work, we evaluated the Rayleigh backscattering coefficients and optical losses in several specialty SMFs and compared them particularly to standard SMFs. Additionally, we investigated the acoustic signal-to-noise ratios of these fibers in a DAS system and discussed their potential to improve DAS system performance for specific applications, such as shorter distances spanning several kilometers.

RC Integrated Systems LLC (RISL) has developed a Fiber Optic Multimodal Sensing (FOMS) System that measures temperature, strain, and pressure in harsh environments. The system uses dual-wavelength fiber Bragg gratings (DW-FBGs) in the side cores of a multicore fiber (MCF) and an extrinsic Fabry-Perot interferometer (EFPI) in the center core. The MCF is coated with high-temperature-resistant (HTR) materials like gold, allowing operation at high temperatures. The FOMS system includes an FBG array, EFPI sensor, a Sensor Interrogation Unit (SIU), and software. For silica fiber with a thermal expansion coefficient of 3×10⁻⁶/°C, the modeled temperature sensitivity is about 17 pm/°C up to 1500°C, and strain sensitivity exceeds 1 pm/με over -5000 to 5000 με. EFPI sensor modeling shows that adjusting the diaphragm thickness and pressure-sensitive area alters the sensor’s sensitivity and range. A 16.8 μm-thick diaphragm with a 500 μm sensitive area provides 2.661 nm/psi sensitivity over 1500 psi, while a 36.2 μm-thick diaphragm with the same sensitive area provides 0.266 nm/psi sensitivity over 15,000 psi.

RC Integrated Systems LLC (RISL) has developed a Fiber Optic Multimodal Sensing (FOMS) System capable of simultaneously measuring multiple parameters, including temperature, strain, and pressure, in harsh environments. The FOMS sensors are based on pairs of dual-wavelength fiber Bragg gratings (DW-FBGs) fabricated in two side cores of a multicore fiber (MCF), along with an extrinsic Fabry-Perot interferometer (EFPI) fiber optic sensor in the center core at the fiber's end. The MCF surface is coated with high-temperature-resistant (HTR) material, such as gold, enabling sensor operation at very high temperatures. The FOMS system consists of an FBG array/EFPI sensor, a Sensor Interrogation Unit (SIU), and supporting software. This paper focuses on the fabrication and initial testing of the FOMS system. The DW-FBGs were tested for temperature measurement, achieving sensitivities of 14.7 pm/°C and 14.8 pm/°C for initial FBG center wavelengths of approximately 1536 nm and 1560 nm, respectively. The strain sensitivity of the DW-FBGs was measured at about 3.5 pm/με and 3.7 pm/με respectively for the two FBGs, and the EFPI pressure sensor showed a sensitivity of approximately -3 mV/psi.