Computational Structural Mechanics Assessment of Large-Scale Fiber Optic Sensor Networks on Highway Bridges for Evaluating Damage Detection Algorithms

摘要

The development of fiber optic strain sensors based on Bragg gratings makes it feasible to install large numbers of sensors on a structure to collect data for health monitoring. The optimum spacing of the sensors in a kilosensor, or megasensor, scale network however depends on the effectiveness of the analysis techniques that will be applied to the data. To investigate this issue, a nonlinear 3-dimensional finite element model, based on LS-DYNA code, has been used to model the dynamics of typical highway bridge structures on a fine spatial scale. Simulations have been run for the structures in an undamaged state and in a damaged state with simulated cracks introduced at specific locations. The resulting response data sets are then sampled at different spatial scales to simulate sensor networks with different spacings. These data sets are used to evaluate the sensitivity of various nonlinear dynamics damage detection analysis techniques. To date, two techniques have been investigated: an energy index approach based on the generalized J contour integral; and a chaos theory method using the Lyapunov spectrum. Both techniques were successful in detecting cracks. For the sample bridge investigated, the required spacing for the J integral method was on the order of 0.3 m. The Lyapunov spectrum method can work on a coarser grid up to 2 m.