Real-time temperature mapping that solves local overheating problems is important for obtaining an optimized thermal design for high-efficiency power transformers. Internal temperature monitoring of operating power transformers can also be leveraged for asset monitoring applications targeting at-fault detection enabling condition-based maintenance programs. However, transformers present a variety of challenging sensing environments such as high-levels of electromagnetic interference and limited space for conventional sensing systems in which to operate. Immersion of some power transformers in insulation oils for thermal management during operation and the presence of relatively large and time varying electrical and occasional magnetic fields make sensing technologies requiring electrical wires or active power at sensing locations highly undesirable. In this work, we investigate dynamic thermal response of a standard single-mode optical fiber instrumented on compact transformer cores by using an optical frequency-domain reflectometry scheme. Correlation between conventional temperature sensing methods and fiber-optic sensing results as well as trade-offs between spatial resolution and temperature measurement accuracy is discussed and spatially resolved real-time monitoring of temperatures in energized transformers is demonstrated.
Monitoring the dissolved gases and oil temperatures are the most responsive and dependable strategies of assessing the running state and health of power transformers. Novel fiber optic sensor approaches that minimize the footprint and cost of chemical sensors compatible with electrical asset insulation oil monitoring can ultimately allow the deployment on a broader range of power assets. Fiber optic sensor arrays based on nanocomposite thin films for simultaneous gas and temperature sensing at low temperatures have been fabricated. The performance of the sensor array was evaluated with two sensing elements comprised of Pd and Au nanoparticle incorporated SiO<sub>2</sub> thin films, which were deposited onto a coreless fiber by dip-coating in series and then fusion-spliced with two multimode fibers on either end. Such fiber optic sensor array showed monotonic responses over a wide range of H<sub>2</sub> at ambient conditions. A response of the localized surface plasmon resonance absorption peak of Au to changing temperatures from ambient to 110 °C was observed. Optical responses of H2 and temperatures showed opposite directions in transmission signals, which were evaluated by using principal component analysis (PCA). The resulting PCA score plot clearly shows discrimination between the characteristic signal of H<sub>2</sub> gas and temperature. In order to realize potential field deployment, the optical testing system employed low-cost components such as LEDs as fixedwavelength light sources and silicon photodiodes as detectors. The potential applications of such sensor arrays are dual-purpose gas and temperature sensing for on-site deployment in power transformers and other grid asset health monitoring.
In this work, a low-cost multipoint fiber optic sensor system for real-time monitoring of the temperature distribution on transformer cores was demonstrated. The temperature sensors are based on multi-mode random air hole fibers infiltrated with CdSe/ZnS quantum dots. Quantum dots resided in multi-mode random-hole core regions can be optically excited by guided UV light with extremely high quantum efficiency. The photoluminescence intensity dependence on the ambient temperatures were used to gauge the local operational temperature of transformer under strong magnetic fields. Multiplepoint temperature sensing systems were developed by bundling quantum dots infiltrated random air hole fibers together. Using a low-cost UV diode laser as a light source and a CCD camera as detector, hundreds of fiber sensors can be interrogated at low cost. This multi-point fiber sensor system, which is free from electromagnetic interference, was used to monitor temperature fluctuation of transformer from the room temperature up to 96°C with better than 1°C accuracy. The proposed fiber optic sensing scheme could overcome the shortcomings of traditional electric sensors and provide a versatile and low-cost approach to map the temperature distribution of electric power systems such as transformers operated in strong electromagnetic fields.
Real-time temperature mapping is important to offer an optimized thermal design of efficient power transformers by solving local overheating problems. In addition, internal temperature monitoring of power transformers in operation can be leveraged for asset monitoring applications targeted at fault detection to enable condition based maintenance programs. However transformers present a variety of challenging environments such as high levels of electromagnetic interference and limited space for conventional sensing systems to operate. Immersion of some power transformers within insulation oils for thermal management during operation and the presence of relatively large and time varying electrical and magnetic fields in some cases also make sensing and measurement technologies that require electrical wires or active power at the sensing location highly undesirable. In this work, we investigate the dynamic thermal response of standard single-mode optical fiber instrumented on a compact transformer core by using an optical frequency-domain reflectometry scheme, and the spatially resolved on-line monitoring of transformer core temperature rise has been successfully demonstrated. It is found that spectral shifts of the fiber-optic sensor induced by the temperature rises are strongly related to the locations inside the transformer as would be expected. Correlation between thermal behavior of the transformer core as derived from standard IR-based thermal imaging cameras and fiber-optic sensing results is also discussed. The proposed method can easily be extended to cover situations in which high accuracy and high spatial resolution thermal surveillance are required, and offers the potential for unprecedented optimization of magnetic core designs for power transformer applications as well as a novel approach to power transformer asset monitoring.
Single crystal fibers like those made from sapphire are capable of operating at higher temperatures than conventional
silica-glass-based fibers. This work aims to construct single-crystal optical fiber sensors capable of providing
environmental data in combustion, high-temperature chemical processing, or power generation applications where
temperatures exceed 1000 °C and standard silica fibers cease to provide useful information. Here, we explore the
functionalization of single crystal fibers using methodologies intrinsic to the crystal growth process or with methods
which do not severely reduce their operating temperature range. While operating a laser-heated pedestal growth system
to produce single-crystal optical fibers from rod feedstock, we continuously vary parameters such as fiber diameter to
produce novel single-crystal linear distributed-sensing devices. The spectral characteristics of those modified devices,
along with sensing performance in a high-temperature harsh-environment are reported. Finally, a technique for
increasing the intrinsic Rayleigh backscattering using femtosecond laser irradiation is discussed for temperature sensing