Thesis (M.S.)--Florida State University, 2004. / Advisor: Dr. John O. Sobanjo, Florida State University, College of Engineering, Dept. of Civil Engineering. Title and description from dissertation home page (viewed June 16, 2004). Includes bibliographical references.

The research presented in this dissertation focuses on the material characterization of ultrahigh performance fiber reinforced concrete (UHP-FRC) at both the microscopic and macroscopic scales. The macroscopic mechanical properties of this material are highly related to the orientation of the steel fibers distributed within the matrix. However, the fiber orientation distribution has been confirmed to be anisotropic based on the flow-casting process. The orientation factor and probability density function (PDF) of the crossing fiber (fibers crossing a cutting plane) orientation was obtained based on theoretical derivations and numerical simulations with respect to different levels of anisotropy and cut planes oriented arbitrarily in space. The level of anisotropy can be calibrated based on image analysis on cut sections from hardened UHP-FRC prisms. Simplified equations provide a framework to predict the mechanical properties based on a single fiber-matrix interaction rule selected from existing theoretical models. Along with the investigation of the impacts from different curing methods and available post-cracking models, a versatile parameterized uniaxial stress-strain constitutive model was developed and calibrated. The constitutive model was implemented in a finite element analysis software program, and the program was utilized in the preliminary design of moveable bridge deck panels made of passively reinforced UHP-FRC. This deck system was among the several alternatives to replace the problematic steel grid decks currently in use. Based on experimental investigations of the deck panels, failure occurred largely in shear rather than flexure during bending tests. However, this shear failure is not abrupt and usually involves large deformation, large sectional rotation, and wide shear cracks before loss of load-carrying capacity. This particular shear failure mode observed was further investigated numerically and experimentally. Three-dimensional FEM models with the ability to reflect the interaction between rebar and concrete were created in a commercial FEM software to investigate the load transfer mechanism before and after bond failure. Small-scale passively reinforced prisms were tested to verify the conclusions drawn from simulation results. In an effort to improve the original design, several shear-strengthened deck panels were tested and evaluated for effectiveness. Finally, methods and equations to predict the ultimate shear capacity iii were calibrated. A two-dimensional frame element based complete moveable bridge finite element model was built for observation of bridge system performance. The model contained the option to substitute any available deck system based on a subset of pre-calibrated parameters specific to each deck type. These alternative deck systems include an aluminum bridge deck system and a glass fiber reinforced plastic (GFRP) deck system. All three alternatives and the original steel grid deck system were evaluated based on the global responses of the moveable bridge, and the advantages and disadvantages of adopting the UHP-FRC deck system are quantified.

Structural Health Monitoring (SHM) is the sensing and analysis of a structure to detect abnormal behavior, damage and deterioration during regular operations as well as under extreme loadings. SHM is designed to provide objective information for decision-making on safety and serviceability. This research focuses on the SHM of bridges by developing and integrating novel methods and techniques using sensor networks, computer vision, modeling for damage indices and statistical approaches. Effective use of traffic video synchronized with sensor measurements for decision-making is demonstrated. First, some of the computer vision methods and how they can be used for bridge monitoring are presented along with the most common issues and some practical solutions. Second, a conceptual damage index (Unit Influence Line) is formulated using synchronized computer images and sensor data for tracking the structural response under various load conditions. Third, a new index, Nd , is formulated and demonstrated to more effectively identify, localize and quantify damage. Commonly observed damage conditions on real bridges are simulated on a laboratory model for the demonstration of the computer vision method, UIL and the new index. This new method and the index, which are based on outlier detection from the UIL population, can very effectively handle large sets of monitoring data. The methods and techniques are demonstrated on the laboratory model for damage detection and all damage scenarios are identified successfully. Finally, the application of the proposed methods on a real life structure, which has a monitoring system, is presented. It is shown that these methods can be used efficiently for applications such as damage detection and load rating for decision-making. The results from this monitoring project on a movable bridge are demonstrated and presented along with the conclusions and recommendations for future work.