Establishing a baseline damage index for reliable damage detection: Full scale validation

Sriram Sivaram Krishnan, Purdue University

Abstract

There is a need to develop effective, fast, automatic and cost-effective structural health monitoring system in the near future to assist in managing our aging infrastructure. The concept of damage is not meaningful without a comparison between two different states of the system, one of which is assumed to represent the initial, and often undamaged state. In general, the modal properties estimated from the multiple sets of data from a real structure show variation in its estimates due to ambient noise, environmental variation and nonlinearities. Hence, it is important to differentiate the change in properties due to damage and environmental variables to demonstrate the robustness of the damage detection method. A drawback of most damage detection methods is that when two sets of acceptable information is passed to a global damage detection algorithm, output damage indices for the structure is obtained without providing information about the acceptability of the results obtained. To detect damage reliably in the presence of these variations, a baseline damage index needs to be established to quantify the change in the damage index due to noise in the structure. It is demonstrated that for a certain noise level in the data, when the damage indices are above a baseline damage index, the damage localization results can be trusted. A graphical representation is used to provide a background for the proposed method. Two damage detection methods: the Angle-between-String-and-Horizontal (ASH) flexibility method and Axial Strain (AS) flexibility method are used to demonstrate the baseline damage index concept. Numerical simulation of a one dimensional beam model and three dimensional truss is used to demonstrate the development of baseline damage index and demonstrate the method. A numerical benchmarking of damage detection method is also proposed using the same concept. Tests conducted on a 17.5-m long experimentally full-scale highway sign support truss were used to demonstrate the effectiveness of the proposed method. A series of detailed tests was used to fully characterize the vibration response of the truss in three orthogonal directions at 44 locations on the structure. Multiple input excitations such as white noise excitation using a shaker to simulate ambient vibration and hammer excitation in multiple directions were used to excite the structure. Five damage cases are studied that replicate scenarios observed during inspection of trusses in the field. A multiple damage case has also been performed in the study. The results demonstrated successfully the applicability of the baseline damage index for reliable damage detection.

Degree

M.S.M.E.

Advisors

Dyke, Purdue University.

Subject Area

Civil engineering|Mechanical engineering

COinS