Fiber Bragg Grating Parameter Synthesis and Sensing Applications Using Genetic Algorithm

• Main Authors: Hsu-Chih Cheng, 鄭旭志
• Other Authors: Jen-Fa Huang
• Format: Others
• Language: en_US
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/74759556938580795520

博士 === 國立成功大學 === 電機工程學系碩博士班 === 94 === This dissertation addresses a wide range of theoretical and experimental issues relating to the parameter synthesis and sensing applications of fiber Bragg gratings (FBGs). Specifically, the dissertation develops novel methods for: (1) reconstructing the multiple parameters of an FBG using a genetic algorithm (GA), (2) sensing arbitrary strain distributions using a GA to inversely search two FBG reflection intensity spectra, and (3) sensing arbitrary strain and temperature distributions simultaneously using a modified GA to inversely search the intensity spectra of FBGs. FBG applications generally require the parameters of the FBG to be known. Many methods have been proposed for reconstructing an FBG’s parameters, including the layer-peeling technique, Fourier transformation, optimization schemes, and so on. However, each method is generally better suited to the parameter synthesis of one particular type of FBG. Accordingly, this dissertation presents a novel, more general, method for the reconstruction of FBG parameters. Reconstructing the complete characteristics of an FBG requires its complex reflection spectrum to be known. Although the reflection intensity spectrum of an FBG is easily obtained using simple instrumentation such as optical power meters or optical spectrum analyzers, obtaining the phase response of the FBG is more difficult. Common methods for obtaining the phase response include the conventional phase-shift technique, interferometric approaches, phase reconstruction algorithms, and so forth. Although each method has certain advantages, they are all limited in some regard when applied in practice. Therefore, a principal objective of this dissertation is to develop a practical approach for obtaining a complete characterization of FBGs of various types. In the proposed approach, the complete set of FBG parameters, namely the grating position, grating period, grating length and index of modulation along the grating length, are extracted from two thermally-modulated reflection intensity spectra using a GA. The performance of the proposed method is verified through its application to a number of simulated and experimental problems. FBGs have emerged as an important component in fiber sensor systems in recent years. Fiber gratings enhance the advantages of fiber sensor systems by adding a powerful wavelength multiplexing capacity. However, FBG-based sensing systems are limited in the sense that they typically provide information of the measurand (generally strain and temperature) only at the level of an average value over the length of the grating. In practice, however, the user is generally more interested in obtaining the distribution profile of the measurand over the FBG length rather than a simple average value. In an FBG, different grating sections reflect different wavelengths when the FBG encounters a non-uniform distribution of the measurand. The FBG therefore produces a complicated reflection spectrum since each section of the grating contributes predominantly at its local Bragg wavelength. This non-uniformity characteristic affects both the intensity and the phase spectra of the FBG. Exploiting this characteristic, this dissertation develops two novel FBG-based sensing arrangements. In the first arrangement, arbitrary strain distributions are sensed by using a GA to inversely search two FBG reflection intensity spectra, while in the second, arbitrary strain and temperature distributions are sensed simultaneously by using a modified GA to inversely search the intensity spectra of four FBGs.