Signal processing and computational modeling for optical fiber sensor devices

Abstract

In the last four decades, special interest has been taken in exploring the characteristics of the optical environment for sensing, giving rise to what would now be one of the largest applications of well-known optical fibers, generally employed to transmit data at high rates. Sensing temperature, pressure, liquid level, deformation among other physical parameters employing optical fibers has become not only a growing branch of research but also a business competing with well-established electrical sensors in the industry. Optical fiber sensors have all the inherent characteristics of a fiber optic cable, such as electromagnetic immunity, small size and weight, multiplexing, and so on. These exclusive features have made fiber sensors so versatile as to become a transformative technology by enabling several industrial processes to be carried out with a higher degree of security. Nowadays, there are several types of optical sensors, these include fiber Bragg grating sensors, Fabry-Perot sensors, interferometric sensors, distributed sensors, polarimetric sensors, polymer fiber sensors, and several others. This immense diversity allows these sensors to be applied in biomedicine, military defense systems, structural health monitoring of bridges, oil wells downhole, pipelines, power transmission lines, and so on. However, to operate properly, theses sensors require adequate processing so that the information measured by them is extracted in the best possible way. In this perspective, this thesis describes a set of signal processing techniques developed for these sensors. The contributions described here cover two classes of optical sensors: Raman distributed temperature sensors and interferometric sensors. For the former one, a method to detect and measure small temperature events on a scale of a few centimeters, as well as a theoretical analysis of the receiver these sensors to improve its sensitivity through hardware modifications are presented. Furthermore, for the latter, a linearization technique is developed to improve its linearity and operation range.