PhD Oral Exam - Farinaz Kouhrangiha, Electrical and Computer Engineering

When studying for a doctoral degree (PhD), candidates submit a thesis that provides a critical review of the current state of knowledge of the thesis subject as well as the student’s own contributions to the subject. The distinguishing criterion of doctoral graduate research is a significant and original contribution to knowledge.

Once accepted, the candidate presents the thesis orally. This oral exam is open to the public.

Abstract

There is a vast field of applications for optical fiber sensors. Fiber Bragg Gratings (FBGs) as an optical sensor is commonly used for Structural Health Monitoring (SHM) to detect various physical phenomena affecting the system to assess its structure in a reliable and accurate manner. External forces such as strain, temperature, and vibration on the fiber change the effective refractive index of the FBG causing a shift in the Bragg wavelength. However, due to the cross-sensitivity of the FBG detection, the effects of each individual parameter on the FBG are a non-trivial task. One way to enhance the accuracy is to disentangle the strain from other affecting parameters such as temperature or vibration for recognizing the real stress that is experienced by the structure. Like any other application, specific FBG sensors are desired for the above purposes to monitor the health structure.

In the present work, an Apodized π-Phase Shifted Fiber Bragg Grating (π-PSFBG) sensor is developed to evaluate the performance of this non-uniform FBG for strain, temperature, and vibration sensing simultaneously. π-PSFBG is selected as an optical sensor to enhance the sensitivity of measurements due to the specific accuracy and spectral characteristics of this type of FBG.

To identify the damage inside the disturbed structure, it is important to detect the irregular behavior of the structure and model the effect of anomalies on the sensor. Here, the design and modeling of the sensor signals have been done by solving coupled-mode equations using the transfer matrix to represent the anomalies effect inside the monitored structure from the reflection spectrum of π-PSFBG. The optimum apodization function has been applied to the spectral signal to improve the properties of the sensor spectrum by suppressing the side lobs.

Moreover, the affecting parameters have been separated to determine the real cause of stress. Towards this end, the reference method has been used to compensate for the strain measurements. To isolate the effects of stress on the structure, the performance of an Apodized π-PSFBG to measure the effects of the above individual parameters on FBG has been studied and characterized with high sensitivity. Also, the effect of affecting parameters on the sensor on a single measurement of Bragg wavelength shift has been identified and discriminated by using an Artificial Neural Networks (ANNs) approach. The neural networks are trained to learn the relationship between the reflection spectrum and the external parameters such as strain, temperature, and vibration. Our investigations would not only characterize the performance of an Apodized π-PSFBG to measure the effects of the above individual parameters with high sensitivity but also yield the minimum error in compensation of strain from affecting gauges. The simulation results show that this highly sensitive modeled sensor is capable of detecting simultaneous parameters under different ranges.

To isolate the exact damage location, the grid structure using the random forest methodology is utilized as the last step in this research. In sensing applications, it is very crucial to detect the damage in every direction. Therefore, to assess the damage position and location with high precision in the monitored structure a grid structure based on the random forest algorithm is able to find the exact location of the damage in every direction, independent of the sensor placement, by corresponding FBG sensor.

The goal of this research is to develop a methodology to model a highly sensitive sensor that can detect and identify the affecting parameters on a monitored structure while characterizing the location of these effects accurately. This sensor would be able to compensate for the effect of external parameters and consequently provide the real cause of damage. presented methods are verified through an extensive set of numerical studies and simulations.