Abstract

A methodology is developed to provide a reliable and quantitative structural health monitoring information with emphasis on three properties, namely on locating the anomaly, modeling the anomaly profile, and identifying the damage inside the disturbed structures. Toward this end, a numerical method is developed to reconstruct the anomaly inside the monitored structure from the reflected spectrum of the Bragg gratings that are fabricated into the single mode (SM FBG) or high birefringence fiber (Hi-Bi FBG).
Firstly, the effects of a non-uniform distribution of the transversal load and temperature on the FBG are analyzed and the perturbed reflected spectrum is modeled by introducing the change of the refractive indices and grating period along the fiber by using the transfer matrix formulation method.
Furthermore, an inverse method based on the genetic algorithm (GA) is developed for reconstruction of non-uniform applied anomalies from the perturbed reflected spectrum. The genetic algorithm retrieves the changes in the characteristics of the sensor from the measured spectra information by encoding the refractive index or the grating period distribution along the Bragg grating into the genes. Moreover, the effects of the simultaneously applied transversal and longitudinal forces on an FBG sensor are analyzed. The study on the effects of the simultaneous transversal and longitudinal forces on an FBG sensor would eliminate the need for the FBGs to be installed on both the orthogonal directions on top of the monitored surface. Consequently, the applied strain measurements can be achieved by parallel fibers in one direction. This will reduce the number of sensors and the complexity of the monitoring system. The perturbed reflected spectra are modeled by the transfer matrix formulation method. Furthermore, the anomaly gradients along the sensor’s length are determined from the intensity spectrum of the sensor by means of the GA. Additionally, the presented functionality of the GA algorithm is tested on a multiplexed FBG sensor system and the anomaly is modeled along a series of the sensors. Consequently, both the location and the model of the anomaly distribution are obtained. Secondly, the effects of the strain and the temperature changes on the Hi-Bi FBG are studied and the reflected intensity spectrums of the polarized modes of the sensor, which are affected by a non-uniform distribution of the temperature or the strain, are modeled theoretically. Each Bragg reflection corresponding to the principal axes of the fiber has different dependencies on temperature and strain. Using this property, the type of the anomaly can be specifically identified and specified. Furthermore, the temperature and the transversal load gradients along the sensor’s length are determined from the intensity spectrum of the sensor by means of the GA. The solution of the genetic algorithm is expressed in terms of the characteristic changes of the sensor, which are correlated with a non-uniform anomaly distribution inside the monitored structure. The presented methods are verified through extensive set of numerical case studies and scenarios.