Parameters Identification Methodology For Flexural Member With Macro-strain Measurement From FBG Sensors

Many civil infrastructures are now deteriorating due to aging, misuse, lacking proper maintenance, and, in some cases, overstressing as a result of increasing load demands and changing environments. Therefore, it's critical to evaluate their current reliability, performance, and condition for the prevention of potential catastrophic events. As the fundament of establishing a health monitoring system for civil structures, damage identification and condition assessment, Structural parameter identification has great significance on the maintenance of structure's safety and normal operations.In the past years, great achievements in long-guage Fiber Bragg Grating (FBG) sensing has been made and it is easier than before to make the vibration-induced macro-strain measurement. Because of the ability to approximate arbitrary continuous functions, artificial neural networks (ANN) have drawn considerable attention in civil engineering for identification in a non-parametric manner. On the other hand structure modal parameters are the function of physical parameters of structures, and damage can be detected by modal information. In this paper, two structural identification methods are proposed using strain measurement from FBG sensors with neural network and strain modal analysis, respectively. The effectiveness and the feasibility of both methods are verified through a small steel test. The contents of this thesis include:1. Based on the brief review on the development of structural identification methods, the background, and main content of this study were introduced. Then the basic theory of artificial neural networks and its applications in civil engineering, as well as fiber-optic sensing techniques and its application in the practice of health monitoring were introduced.2. A direct parameter identification method was proposed based on the macro-strain time series response of structure and BP neural network. First the fundament of the proposed methodology was explained, and the implementation procedure with ANN was described. Test on two steel beam specimens were carried out and the macro-strain measurements were employed to identify the stiffness and damage of the steel specimen with proposed ANN based identification method.3. A macro-strain mode based damage identification algorithm is proposed for structural local damage identification with strain measurements from long-gage FBG sensors. A damage indexβwas defined using the changes of the structural modal macro-strains before and after damage. Based on the relationship between the damage index and the damage extent, the stiffness and damage of the small-scale beam were identified with acceptable accuracy.The results show that the proposed two methods can identify structural parameters and damage using macro-strain measurement from FBG sensors for engineering structures.