Embedding fiber optic sensors in critical components is a key step for real-time monitoring of structural conditions during service and supporting autonomous system operations. Successful integration of these sensors necessitates effective interfacial bonding between the fiber and matrix, good integrity and functionality of the embedded sensors, robust mechanical strength of the matrix materials, and the ability to retain these properties during transient thermal and stress events. This study demonstrates the encapsulation of fused silica optical fibers in stainless steel and nickel through the electric field-assisted sintering (EFAS) process. Copper-coated and gold-coated single mode optical fibers were embedded under different EFAS conditions. The resulting components with embedded sensors were evaluated using advanced microscopy and optical frequency domain reflectometry (OFDR) to assess the aforementioned critical aspects of embedding. The results indicate that both copper- and gold-coated fibers can be successfully embedded in stainless steel and nickel with good fiber integrity and fiber-matrix bonding. Samples fabricated under optimal conditions passed helium leak testing, confirming effective interfacial bonding. Microstructural characterization revealed excellent fiber-matrix adhesion and interdiffusion of elements across the interface. The functionality of the embedded fibers was evaluated through OFDR scans, which revealed signal insertion loss of 0.43 – 0.52 dB for nickel samples and 0 – 0.75 dB for stainless steel samples at the embedding sites. Additionally, the embedded fibers underwent cyclic thermal treatment between 500°C and 700°C. The fibers maintained good integrity and interfacial characteristics, demonstrating their ability to survive cyclic thermal events for sensing in harsh environments.

To enhance the prediction accuracy and efficiency of the wireless outdoor heatmap, we propose a novel federated Gaussian Process (GP) approach combined with Bayesian Model Averaging (BMA). Traditional centralized GP models need extensive communication between distributed sensors and a central server, which leads to inefficiencies, increased computational costs, and potential privacy risks. Additionally, the GP model's log-likelihood function is optimized using the entire dataset, which makes it incompatible with standard federated aggregation techniques such as Federated Averaging (FedAvg). To overcome these challenges, our approach enables each sensor to process its data locally with GP algorithms by mapping Received Signal Strength (RSS) to corresponding locations. At the central server, BMA predicts pseudo-labels from limited global data to create a pseudo-labeled set for knowledge distillation. This allows the central server to train a global student GP model that updates parameters rather than averaging local models directly as in FedAvg. The global model is then sent back to sensors for further iterations. We evaluate our approach using real-world RSS data from the National Science Foundation (NSF) funded Platform for Open Wireless Data-driven Experimental Research (POWDER) at the University of Utah. The experiment results demonstrate that the proposed federated GP model significantly outperforms existing methods, including the federated GP schemes using FedAvg and classical Alternating Direction of Multipliers Method (cADMM) and federated Neural Network (NN)-based scheme.

We present a novel Bayesian learning approach to outdoor radio heatmap construction utilizing deep Gaussian Process (GP). The proposed approach employs a two-layer hierarchy that is capable of modeling more complex input-output relations than the standard single-layer GP. Since deriving the exact model likelihood is challenging, a lower bound is optimized instead to find the optimal model parameters. Typically, inducing points are used to facilitate low-rank approximation of covariance (kernel) matrices for computation speedup. However, the inaccuracy induced by inducing points can accumulate when stacking multiple layers of GP which may degrade the performance of deep GP. Moreover, since inducing points need to be learned, having them at all layers of deep GP also incurs computational burden. To overcome the above challenges, in contrast to the canonical deep GP model, we use a modified architecture where a full standard GP resides in the first layer and inducing points are only introduced for the second layer. This modified architecture strikes a balance between model accuracy and training complexity. In the proposed model, the noise parameter of the first GP layer is also eliminated to improve the training efficiency as the noise parameter at the output of the second layer suffices to model the uncertainty in the output. The proposed approach is evaluated on real-world datasets, collected from the Platform for Open Wireless Data-driven Experimental Research (POWDER) located at the campus of the University of Utah. Experimental results show that the proposed approach can achieve superior performance on various training and testing data configurations compared to canonical deep GP schemes, DNN-based and GP-based methods.

Surface acoustic wave (SAW) devices are a subclass of micro-electromechanical systems (MEMS) that generate an acoustic emission when electrically stimulated. These transducers also work as detectors, converting surface strain into readable electrical signals. Physical properties of the generated SAW are material dependent and influenced by external factors like temperature. By monitoring temperature-dependent scattering parameters a SAW device can function as a thermometer to elucidate substrate temperature. Traditional fabrication of SAW sensors requires labor- and cost- intensive subtractive processes that produce large volumes of hazardous waste. This study utilizes an innovative aerosol jet printer to directly write consistent, high-resolution, silver comb electrodes onto a Y-cut LiNbO 3 substrate. The printed, two-port, 20 MHz SAW sensor exhibited excellent linearity and repeatability while being verified as a thermometer from 25 to 200 ∘ C. Sensitivities of the printed SAW thermometer are -96.9× 10^-6^∘ C −1 and -92.0× 10^-6^∘ C −1 when operating in pulse-echo mode and pulse-receiver mode, respectively. These results highlight a repeatable path to the additive fabrication of compact high-frequency SAW thermometers.

The goal of this work was to investigate the in-core performance of sapphire optical fiber temperature sensors and to develop clad sapphire optical fibers for in-core instrumentation.We fabricated clad sapphire optical fibers and evaluated the distributed sensing performance of these sensors via optical backscatter reflectometry under high fluence and combined radiation and temperature effects.A series of irradiations was completed to evaluate the effect of irradiation on sapphire optical fiber temperature sensors and to determine the operational limits of these sensors.• Objective 1: Fabricate sapphire optical fiber sensors.• Objective 2: Evaluate the clad sapphire fiber to verify single-mode behavior and determine and characterize the light modes supported by optical fibers.• Objective 3: Characterize the in-core temperature sensing of sapphire optical fiber, as well as the combined temperature and irradiation effects.• Objective 4: Evaluate the lifetime and performance of the sensor under irradiation to high neutron fluence.Objectives 1, 2, and 3 were completed during the first 2 years of the project.Due to the Covid pandemic, Objective 4, a high-fluence irradiation performed at the Massachusetts Institute of Technology Research Reactor (MITR), was delayed, as partner facilities were subject to mandatory shutdowns and required a 1 year, no-cost extension.This irradiation was eventually completed on December 12, 2022.This work indicates that sapphire optical fiber sensors may be a solution for ultra-high-temperature applications in which traditional silica optical fibers are prone to fail.Sapphire sensors are potentially suitable for experiments featuring temperatures above 700℃ for long periods of time, or for any length of time above 1000℃.Experiments featuring a low total fluence, such as irradiations conducted in the Transient Reactor Test (TREAT) facility, also represent good applications for sapphire optical sensors.Additional work is required to characterize the sapphire fiber cladding performance, which falls outside the scope of this project, as well as the effects of high temperatures on the response of the fiber.A comprehensive material study is recommended as future work to evaluate the attenuation in sapphire under irradiation, and how that attenuation changes with irradiation temperature.The drift and attenuation in the fiber at temperatures of up to 1600℃ and a total fluence of up to 2.9 x 10 17 n/cm 2 was minimal, and the fibers returned to baseline after being heated to 1600℃ under irradiation.This is promising for the future use of sapphire optical fibers in advanced reactors.

In-situ measurement of surface acoustic wave (SAW) resonators were used to characterize piezoelectric materials behavior in a nuclear reactor environment. Lithium niobate and aluminum nitride devices were tested up to 500 C temperature and 450 kW reactor power (1.9e12 n/cm2s). The resonant frequency responds to step changes in temperature and power. Materials’ response are inferred from the observed resonant frequency changes, in particular the kinetic behavior is used to determine mechanisms responsible for observed changes. The observed response is attributed to two mechanisms: temperature increase from gamma heating and accumulation of radiation-induced defects. Both of these mechanisms alter the physical properties of the piezoelectric materials, particularly the elastic constants. The demonstrated repeatability and reliability of SAW devices are attractive for sensor applications in extreme environments.

An experimental system was designed and optimized to acquire real-time electrical resistance of the monitors during annealing. The system was then used to anneal irradiated silicon carbide (SiC) monitors. The altered resistance of the SiC monitors between the first heating cycle and the mean of the heating and cooling cycles that followed was found to significantly change when the annealing temperature exceeded the peak irradiation temperature. This was validated using nine irradiated SiC monitors annealed over two heating and cooling cycles. Of these SiC monitors, three were annealed for the first time. The remaining six monitors were already annealed, but four of these were found to contain residual defects that resulted in reasonable estimates of peak irradiation temperatures. Those estimated peak irradiation temperatures were statistically indifferent from the manual isochronal annealing method temperatures. The results demonstrate a preliminary potential for the two-cycle approach to replace or augment the current manual post-irradiation examination (PIE) process for extracting SiC monitor peak irradiation temperature.