Structural health monitoring and damage identification using best approximated dead load stress redistribution.

摘要

Faced with an aging infrastructure system in the United States, a structural health monitoring system capable of automatically detecting, locating and assessing structural deterioration or damage is highly desired. A novel best approximation method for damage identification in large civil structures based on measuring changes in the static responses of the structures due to dead load has been developed. Damage in a frame structure is modeled by a hinge plus a rotational spring at the damage location. The damage identification approach is based on finding a "null plane", which is a unique identity of each structural member, such that the measurement is closest to its orthogonal projection on this null plane. The detection procedure resembles the process of finding the shortest distance from a point to a plane in elementary geometry. The method is developed in a flexibility formulation and a stiffness formulation. The damage sensitivities of the proposed method subject to measurement noise are studied using the Fisher Information Matrix as well as Monte Carlo Simulation. The damage sensitivities are found highly dependent on the damage location and the sensor locations. To increase the damage sensitivities, sensors are optimally located by maximizing the determinant of Fisher Information Matrix.; Damage identification on single-span beams using experimental data from a laboratory test has been attempted. Structural response data was obtained from tests conducted on aluminum beams both in undamaged and damaged states. Damage identification based on these experimental data yields relatively accurate results. In general, the proposed best approximation method for damage identification is capable of automatically detecting, locating and assessing the structural damage in a simple, fast and passive way. Continuing work must be done to reach the goal of a robust damage identification method that can be instrumented on field-testing.